Ccr4-not ubiquitin ligase signaling regulates ribosomal protein homeostasis and inhibits 40S ribosomal autophagy

The Ccr4-Not complex contains the poorly understood Not4 ubiquitin ligase that functions in transcription, mRNA decay, translation, proteostasis, and endolysosomal nutrient signaling. To gain further insight into the in vivo functions of the ligase, we performed quantitative proteomics in Saccharomyces cerevisiae using yeast cells lacking Not4, or cells overexpressing wild-type Not4 or an inactive Not4 mutant. Herein, we provide evidence that balanced Not4 activity maintains ribosomal protein (RP) homeostasis independent of changes to RP mRNA or known Not4 ribosomal substrates. Intriguingly, we also find that Not4 loss activates 40S ribosomal autophagy independently of canonical Atg7-dependent macroautophagy, indicating that microautophagy is responsible. We previously demonstrated that Ccr4-Not stimulates the target of rapamycin complex 1 (TORC1) signaling, which activates RP expression and inhibits autophagy, by maintaining vacuole V-ATPase H+ pump activity. Importantly, combining Not4 deficient cells with a mutant that blocks vacuole H+ export fully restores RP expression and increases 40S RP autophagy efficiency. In contrast, restoring TORC1 activity alone fails to rescue either process, indicating that Not4 loss disrupts additional endolysosomal functions that regulate RP expression and 40S autophagy. Analysis of the Not4-regulated proteome reveals increases in endolysosomal and autophagy-related factors that functionally interact with Not4 to control RP expression and affect 40S autophagy. Collectively, our data indicate that balanced Ccr4-Not ubiquitin ligase signaling maintains RP homeostasis and inhibits 40S autophagy via the ligase’s emerging role as an endolysosomal regulator.

Cells dynamically respond to nutrient availability and stress through mechanisms that require communication between the gene expression machinery and metabolic-responsive organelles, including the endolysosomal compartment and mitochondria (1,2).This communication allows cells to coordinate their growth and proliferation with favorable nutrient environments, while it has an equally significant role in limiting anabolic and proliferative activity under unfavorable environments.The dysregulation of these processes causes many complex diseases and developmental syndromes, including cancer, neurodevelopmental disorders, aging, and metabolic disease (2).Deciphering these interrelationships remains an unrealized goal as they are inherently complex with considerable functional redundancy.
The Ccr4-Not complex is a large multisubunit complex that is conserved throughout evolution and has critical roles throughout the gene expression process.Ccr4-Not originally was identified in budding yeast where it consists of the Ccr4, Caf1, Caf40, Caf130, and the Not1-5 subunits (3).The complex's best-understood role is the regulation of cytoplasmic mRNA decay via the Ccr4 deadenylase subunit, and Ccr4-Not mediated mRNA degradation is highly conserved in metazoans (3,4).Additionally, Ccr4-Not regulates RNA polymerase II (Pol II)-dependent transcription initiation and elongation, and mRNA nuclear export (3,5,6), while it also controls RNA Polymerase I (Pol I) transcription of ribosomal RNA genes downstream of the target of rapamycin complex 1 (TORC1) signaling pathway (7).Although Ccr4-Not has important transcriptional activities, its mechanistic roles in Pol I and Pol II transcription are less defined compared to its role in mRNA degradation.
While the Ccr4 deadenylase remains the best characterized enzymatic subunit in the complex, Not4 is a highly conserved ubiquitin ligase whose in vivo functions are poorly understood.Not4 has an N-terminal RING domain that interacts with the paralogous Ubc4 and Ubc5 E2 enzymes to mediate direct substrate ubiquitination (8).The best characterized Not4 substrates are the histone demethylase Jhd2 and the Cyclin C subunit of the Mediator complex, and Not4-dependent ubiquitination stimulates their degradation by the proteasome (9,10).Immediately C-terminal to the Not4 RING domain is a highly conserved RNA recognition motif (RRM) and C3H1 motif that are collectively referred to as the RRM-C domain (11), which has homology with other RNA binding proteins.However, while Not4 crosslinks to RNA in vivo (12), whether it binds RNA directly through the RRM-C domain is unknown.Not4 RING mutants alone do not completely inactivate Not4 in vivo, yet combining both RING and RRM-C mutations simultaneously does phenocopy a Not4 gene deletion mutant (not4D) (11).Therefore, both domains are required for Not4 in vivo function, and the combination of RING and RRM-C domains also makes Not4 a completely unique ubiquitin ligase in the eukaryotic proteome (13).Besides direct substrate ubiquitination, Not4 also behaves as an adaptor via recruitment of the HECT-domain ligase Rsp5/NEDD4 (14).This capacity to recruit Rsp5/NEDD4 has the potential to expand Ccr4-Not's ubiquitin signaling role into areas previously not associated with Ccr4-Not.For example, while Rsp5 regulates transcription and mRNA export (15)(16)(17), it also controls protein sorting and degradation through the endolysosomal compartment and it regulates autophagy (18)(19)(20)(21)(22)(23).
The effect individual Ccr4-Not subunits have on the transcriptome has been described (30,(39)(40)(41), yet the impact that Not4 ligase signaling has on the proteome remains unclear.How the Ccr4-Not ligase recognizes substrates for ubiquitination is unknown, and only a few substrates have been identified.To address this deficiency, we have performed tandem mass tag (TMT) labeling and quantitative mass spectrometry proteome analysis using not4D reconstituted with a control vector, or vectors overexpressing wild-type (WT) Not4 or an inactive Not4 mutant that phenocopies not4D (11).The underlying rationale for this approach was to identify proteins whose expression is repressed by Not4 but not by the Not4 mutant, which we have identified and reported in this study.This work has led to the unexpected discovery that both Not4 ligase inactivation and Not4 overexpression inhibit the expression of ribosomal protein (RP) and ribosomal biogenesis (Ribi) factors, which have important functional consequences for protein synthesis and proteostasis.Additionally, we find that the Ccr4-Not ligase actively represses 40S ribosome autophagy, and we provide genetic evidence that suggests both the RP repression and 40S autophagy activation is related functionally to alterations in the endolysosomal pathway.

Both increased Not4 expression and loss of its ligase activity inhibit RP and Ribi protein expression
Although Ccr4-Not regulation of the transcriptome has been defined in-depth (39,40), the impact that the Ccr4-Not ligase has on the proteome remains incompletely defined.Furthermore, most Ccr4-Not ligase studies have used cells either lacking Not4 expression entirely (not4D) (42), express Not4 partial deletion constructs (14), or have reconstituted not4D with Not4 RING mutants that do not completely phenocopy not4D (8).Accurate assessment of Not4 function necessitates using a full-length Not4 ligase that incorporates into Ccr4-Not yet a functionally inactive ligase to avoid indirect effects on Ccr4-Not activity independent of its ubiquitin ligase role.Thus, we transformed wild-type (WT) and not4D with a control vector, vector overexpressing C-terminal FLAG-tagged WT Not4 (NOT4), or a Not4 mutant (NOT4RR) that contains mutations in both the RING and RRM-C domains.We previously demonstrated that both Not4 and Not4RR incorporate into Ccr4-Not, yet Not4RR is functionally null and phenocopies not4D (11).Plasmid expressed NOT4 restored all not4D phenotypes, including sensitivity to heat stress, while NOT4RR failed to restore any not4D phenotype tested (Fig. 1A and see below) (11).Importantly, plasmid expressed NOT4 and NOT4RR transcripts are expressed at higher levels relative to NOT4 expressed from its endogenous chromosomal locus (Fig. 1B), and both Not4 and Not4RR are readily detectable by anti-FLAG immunoblot (IB) (Fig. 1C).We reasoned this Not4 and Not4RR overexpression system could be used to identify proteins whose abundance is sensitive to Not4 ligase activity and/or expression levels.These proteins then could be pursued as possible novel Ccr4-Not ligase substrates.
To identify the differentially expressed proteins in these conditions, whole-cell extracts (WCEs) from at least three biological replicates from Figure 1A were prepared and subjected to LC-MS-MS analysis using Reporter Ions Quantification approach based on Tandem Mass Tag (TMTpro)labeling.Quantifiable results for over 3300 proteins in each condition were achieved, and principal component analysis (PCA) revealed high reproducibility between the independent replicates as evidenced by their clustering (Fig. 1D).Intriguingly, while not4D + NOT4 rescues all not4D phenotypes (Fig. 1A) (11), it clusters distinctly from the WT + vector control, which indicates Not4 overexpression does alter the proteome (Fig. 1D).The not4D + vector and NOT4RR samples The Ccr4-Not ligase inhibits 40S ribosome autophagy clustered similarly (Fig. 1D), which is predicted since the Not4RR mutant phenocopies not4D (Fig. 1A) (11).To identify the differentially expressed proteins (DEPs) between the experimental conditions, the data were analyzed using a 1.5fold change in expression with an FDR<0.05relative to the WT + vector control.Using these criteria, 959 (not4D + vector), 290 (not4D + NOT4), and 701 (not4D + NOT4RR) DEPs were identified (Fig. 1E).Venn analysis of these DEPs identified a substantial overlap of 214 proteins between the vector, NOT4, and NOT4RR expressing cells, thus indicating a protein subset whose expression is sensitive to both the Not4 expression level and its ligase activity (Fig. 1E).Pairwise analyses also reveal exclusive overlap between vector and NOT4 (26 proteins), NOT4 and NOT4RR (9 proteins), and vector and Each experimental condition had four independent biological replicates except for not4D + NOT4, which had three independent biological replicates.E, the differentially expressed proteins (DEPs) were identified by using a 1.5-fold change in expression and an FDR < 0.05.The resulting DEPs then were subjected to Venn analysis.The total number of DEPs in each condition relative to the WT + vector control is 959 (not4D + vector), 290 (not4D + NOT4), and 701 (not4D + NOT4RR).
NOT4RR (422 proteins) (Fig. 1E).Additionally, the not4D + vector, NOT4, or NOT4RR cells exhibit condition-specific DEPs, with 297 DEPs in the not4D vector samples, while 41 and 56 DEPs were specific for the NOT4 and NOT4RR samples, respectively (Fig. 1E).These data indicate that both Not4 expression and its ligase activity have overlapping as well as selective proteome effects.Importantly, we find that the not4D + vector exhibits reduced expression (greater than 1.5fold change) of the Caf40 and Not2 Ccr4-Not subunits compared to the WT + vector control while reconstituting cells with either Not4 or the Not4RR mutant restores Caf40 and Not2 protein expression (less than 1.5-fold change for either) (Fig. S1).These data are consistent with the observation that Ccr4-Not subunit loss can destabilize additional complex members (43).While some of the differentially expressed proteins may be due to non-specific effects caused by overexpressing Not4 or the Not4RR mutant, the identified DEPs specifically in Not4 overexpressing cells provide an initial set of factors to test in future efforts to determine if they are candidate Not4 substrates.The protein differences between not4D + vector and not4D + NOT4RR also validate the rationale for performing the proteome analyses in the presence of a full-length, functionally inactive Not4 ligase mutant to avoid altering the Ccr4-Not complex and affecting additional Ccr4-Not activities that are independent of its ubiquitin ligase activity.
The DEPs for each experimental condition were subjected to Gene Ontology (GO) analysis through the STRING database to identify overrepresented functional networks (44).For not4D + vector, the GOs were dominated by categories related to the nucleolus, preribosome, and cytoplasmic ribosome (Fig. 2A).When these DEPs were further sub-divided into up and down DEPs, the down DEPs were specific for the nucleolar and ribosomal-related GOs, while the up DEP GOs were enriched for proteasome regulatory particle, polarized growth, and cytoskeletal functions (Fig. S2).Ribosomal-specific GOs also were overrepresented in not4D + NOT4 expressing cells, although not to the same extent as in not4D + vector.Additionally, a substantial number of the upregulated NOT4 GOs related to mitochondrial function and metabolism, while the NOT4 repressed DEP GOs were specific to ribosomal function (Figs.2B and S2).These data indicate that increased Not4 expression reduces ribosomal factors but enhances the expression of mitochondrial proteins.A similar analysis of the aggregate not4D + NOT4RR DEPs indicated nucleolar, ribosomal, and cytoplasmic stress granule GOs were overrepresented (Fig. 2C).The down DEPs were specific for ribosomal GO categories, while no specific GO categories were overrepresented in the up DEPs (Fig. S2).Collectively, these data indicate that both decreased and increased Not4 activity represses RPs and ribosomal biogenesis (Ribi) factors.
We next examined whether the RP repression was due to inhibition of their mRNA expression.RP genes are coordinately regulated (45), so we chose several RPs from the proteomic dataset and performed RT-qPCR to quantify their mRNA levels.A minor, but statistically significant (p< 0.05 or greater by two-way ANOVA), decrease in RPS5 mRNA was found in the not4D vector, NOT4, and NOT4RR cells relative to the WT + vector control (Fig. S3).However, the expression of the remaining RP mRNAs either was unaffected or upregulated in these cells relative to the WT + vector control (Fig. S3).As such, these data indicate that RP inhibition in the not4D + vector, NOT4, and NOT4RR cells cannot be explained by decreased expression of RP mRNAs.Not4 contributes to translational regulation in part by monoubiquitinating the RP Rps7a and the nascent polypeptide-associated complex factors Egd1 and Egd2 (Egd1/2) to promote co-translational quality control (24,26).The proteome data revealed that Rps7a and Egd1/2 expression was unaffected in all experimental conditions (less than 1.5-fold change in abundance, see Fig. S4), indicating the global RP inhibition is not caused by repression of known Not4 ribosomal substrates.
The recent study from Allen et al. utilizing SILAC analysis reported the proteome differences between WT and not4D cells and found that the expression of a subset of RPs and translation elongation factors increased in not4D (42).We compared the RPs identified by the Allen study (proteins represented by a minimum of three peptides and meeting a 1.5 fold-change in expression with a p-value < 0.05) to the RPs identified in our not4D + vector dataset.A total of 38 proteins from the Allen study was compared to our results (Table S2), with 21 proteins found to be unique to the Allen study, 44 unique to our dataset, and 17 overlapped (Fig. 2D).The overlapping 17 RPs then were compared by plotting their Log 2 expression to compare the directionality of their expression change.Of these 17 proteins, 11 were decreased in both studies while only six (Rpl22a, Rpl37a, Rpl38, Rpp1a, Rps12, and Rps5) were increased in the Allen study but repressed in our dataset (Fig. 2E).Therefore, while our study finds a greater number of RPs to be downregulated in not4D, for the RPs that do overlap only six are altered in an opposing fashion.

Loss of Not4 ligase activity impairs protein synthesis and induces 40S RP degradation
Ccr4-Not mutants, including not4D, reduce polysome formation and alter the relative ratios of 40S and 60S ribosomes (24).To test if the RP repression detected was associated with effects on protein synthesis, the WT control, and the not4D vector, NOT4, and NOT4RR expressing cells were metabolically labeled with puromycin.Puromycin incorporates nascent protein chains to block protein synthesis, and metabolic labeling of cells with puromycin can be detected by a-puromycin IB, which is used as a readout of active protein synthesis (46).A time-dependent increase in puromycin incorporation was readily detectable in the WT + vector control cells, while puromycin incorporation was reduced dramatically in both the not4D + vector and NOT4RR samples (Fig. 3A).Although NOT4 overexpression inhibits a large network of RPs and Ribi factors (Figs. 2 and S5), NOT4 restored protein synthesis to the WT state (Fig. 3A).Therefore, the ribosomal inhibition in NOT4 cells is insufficient to disrupt protein synthesis, while protein synthesis in not4D and NOT4RR expressing cells is profoundly inhibited.
The Ccr4-Not ligase inhibits 40S ribosome autophagy  A-C, bubble blot analysis of the top cellular component gene ontology (GO) categories for the indicated experimental conditions.GO analysis was performed through the STRING database.File S3 has the GO category and individual protein results used to generate these bubble blots.D, Venn analysis of the reduced RPs identified in this study versus the RPs differentially expressed in not4D from the Allen study (42).E, heatmap analysis of the Log 2 expression for each overlapping RP identified in (D).File S3 has the RPs listed used for this analysis.

The Ccr4-Not ligase inhibits 40S ribosome autophagy
The impact Not4 expression and ligase activity have in vivo was probed further by analyzing sensitivity to proteostatic stress (incubation at 37 C and sensitivity to azetidine-2carboxylic acid (AZC) which causes protein folding stress), nutrient stress (the TORC1 inhibitor rapamycin), or translational stress (cycloheximide and Hygromycin B) (11).For an additional comparison, WT cells also were transformed with the NOT4 or NOT4RR expressing plasmids.As previously reported (11), not4D vector and NOT4RR cells exhibited marked sensitivity to the stress conditions tested, while NOT4 rescued all of these phenotypes (Fig. 3B).In WT cells, NOT4RR expression sensitized cells to proteostatic stress (heat stress and AZC), but it had no effect on nutrient stress or cycloheximide-induced translational stress (Fig. 3B).These results indicate that Not4RR functions as a dominant negative in WT cells by competing with endogenous Not4 for Ccr4-Not incorporation, suggesting that Ccr4-Not's role in proteostasis is more sensitive to Not4 ligase activity compared to other stress responses.Intriguingly, while NOT4 had no effect on the sensitivity of WT cells to most stressors, both NOT4 and NOT4RR profoundly inhibited WT growth in the presence of Hygromycin B (Fig. 3B).Hygromycin B stabilizes tRNA interactions with the ribosomal A-site to inhibit tRNA translocation (47), while cycloheximide binds the ribosomal 60S Esite to block ribosome elongation (48).These results suggest that both increased and decreased Not4 ligase activity negatively affects translational control, possibly through mechanisms affecting tRNA-ribosome interactions.Overall, these data suggest that Not4-dependent proteostasis regulation is more dependent on Not4 activity relative to other stress responses and that increased Not4 expression sensitizes cells to mechanistically selective translational inhibitors.
To further delineate Not4-dependent effects on RP expression, we chose two representative RPs, the 40S subunit Rps9a and the 60S subunit Rpl36a, that are contained in the 214 overlapping DEPs for all conditions (Fig. 1D).An EGFP epitope tag was integrated in-frame at their genetic loci in WT, not4D, and ccr4D such that Rps9a and Rpl36a are expressed as in-frame C-terminal GFP fusions from their native promoters.Using these reporters, we found not4D reduces the expression of both RPs, as predicted from the proteomic results, while ccr4D had no effect on either RP (Fig. 3C, short a-GFP exposure).These data confirmed that RP inhibition in not4D is independent of Ccr4-dependent transcription or mRNA deadenylation.While Rps9-GFP is 56 kDa, longer exposure of the a-GFP IB revealed a discrete, faster migrating Rps9a-GFP band in not4D at 40 kDa and a doublet at 27 to 34 kDa in ccr4D (Fig. 3C).These smaller Rps9a-GFP bands are absent from the WT, and smaller RP-specific bands are not present in the WT, not4D, or ccr4D Rpl36a-GFP expressing cells (Fig. 3C).These results suggest that the 40S RP may be partially degraded in Ccr4-Not deficient cells.We confirmed that these GFP tags did not increase the cell's sensitivity to Hygromycin B-induced translational stress (Fig. 3D).In addition, analysis of the WT Rps9a-GFP strain by confocal microscopy demonstrates that Rps9a-GFP is broadly found throughout the cell as predicted (Fig. S6), indicating that the GFP tag does not disrupt Rps9a distribution to cause its degradation.Collectively, these data indicate that these Rps9a-GFP and Rpl36a-GFP effects are not due to the GFP tag causing impairment in ribosome function, but instead, they reflect alterations specifically due to loss of the Not4 ligase.
To confirm the inhibitory effects that Not4 loss of function has on additional RPs and Ribi factors identified in our proteomics, we integrated in-frame C-terminal GFP tags at the genes encoding an additional 40S RP gene (RPS30B), a Ribi factor (SRP40), and a non-ribosomal related spliceosomal factor (PRP11) (49).The not4D repressed full-length Rps30b-GFP expression compared to WT as predicted by our proteomics (Fig. 3E).Although not as pronounced as the not4D Rps9a-GFP degradation product (Fig. 3C), we also detected an Rps30b-GFP degradation product that migrates at approximately the same position (43 kDa) (red arrow, Fig. 3E).To confirm that the not4D 40S RP repression is specific only for those RPs identified in our dataset, we utilized an antibody recognizing the endogenous 40S subunit Rps3 (not repressed in any of the experimental conditions) and confirmed that not4D does not affect its expression (Fig. 3E).Intriguingly, while not4D also represses Srp40-GFP as predicted, we also detected a substantial GFP-specific degradation product at 27 kDa, which is the size of free GFP (Fig. 3F).These partial and/or complete degradation products are specific for the 40S RPs and Ribi factor since not4D represses Prp11-GFP expression, consistent with our proteomic results, but it does not cause Prp11-GFP degradation (Fig. 3G).A recent report demonstrated that ccr4D upregulates basal macroautophagy in repressive (nutrientrich) conditions (32).Yeast autophagy flux is assessed by fusing GFP to candidate autophagy substrates and monitoring the release of free GFP (27 kDa) that occurs when autophagy cargoes are degraded in the vacuole since GFP resists vacuolar degradation (50).The faster migrating Rps9a-GFP, Rps30b-GFP, and Srp40-GFP bands indicate that under these nutrient-rich repressive conditions, Ccr4-Not ligase loss may increase 40S and Ribi, but not 60S, autophagy (referred to collectively as ribophagy) (51).
The effect that Not4 ligase activity and expression level has on Rps9a-GFP and Rpl36a-GFP also was assessed directly by transforming the WT and not4D Rps9a-GFP and Rpl36a-GFP reporter strains with control, NOT4, or NOT4RR expressing plasmids.In addition, a different plasmid was transformed into not4D that expresses NOT4 from its endogenous promoter (NOT4ORF) so that NOT4 expression mirrors that of WT cells.Rps9a-GFP remained repressed in the not4D vector, NOT4, and NOT4RR cells (Fig. 4A), which replicates the results from the proteomic analysis (Fig. 4B), while the NOT4ORF restored Rps9a-GFP back to WT levels (Fig. 4A).Furthermore, the partial Rps9a-GFP degradation product in the not4D vector and NOT4RR samples was absent in both NOT4 and NOT4ORF cells even though NOT4 overexpression does not restore Rps9a-GFP expression (Fig. 4A).In both not4D vector and NOT4 overexpressing cells, Rpl36a-GFP was decreased compared to WT (Fig. 4C), which is consistent with our proteome results (Fig. 4D), while NOT4ORF restored The Ccr4-Not ligase inhibits 40S ribosome autophagy Rpl36a-GFP levels back to the WT state (Fig. 4C).Interestingly, NOT4RR also restored Rpl36a-GFP expression back to WT levels (Fig. 4C), which differs from our proteome results (Fig. 4D).The underlying explanation for this discrepancy currently is unclear.This effect likely stems from an unknown effect the GFP tag has on Rpl36a regulation even though the Rpl36a-GFP tag does not alter sensitivity to ribosome stress (Fig. 3D).Collectively, these data indicate that a narrow range of Not4 ligase activity maintains wild-type RP expression, and these RP effects encompass both 40S and 60S RPs.They also indicate that the partial Rps9a-GFP degradation in not4D specifically results from a lack of Not4 ligase activity independently of Not4-mediated Rps9a repression.

Not4 ligase loss represses 40S RP expression independent of proteasome-mediated degradation but dependent on the endolysosomal pathway
Recent studies determined that loss of Not4-mediated ribosomal Rps7a monoubiquitination causes defects in protein solubility and increases protein aggregate formation (24,25).To test for this possibility, equal numbers of WT and not4D cells expressing Rps9a-GFP or Rpl36a-GFP were lysed by boiling in SDS-PAGE sample buffer and then analyzed by a-GFP IB.Even when samples were prepared under denaturing conditions, not4D repressed both RPs (Fig. 5A).Importantly, overexpressing WT Not4 still inhibits a large network of RP and Ribi factors (Figs. 2 and S5) although it fully restores the The Ccr4-Not ligase inhibits 40S ribosome autophagy The Ccr4-Not ligase inhibits 40S ribosome autophagy protein synthesis defects (Fig. 3A), so the RP inhibition detected cannot be explained by the aggregate formation and protein insolubility.
RPs produced in excess beyond that needed for ribosome biogenesis are flagged with ubiquitin and degraded through the ubiquitin-proteasome (UPS) system (52,53).Using Rps9a as a model RP, we assessed if Rps9a was ubiquitinated in not4D cells to promote RP repression through proteasome-mediated degradation.Initially, WCEs from WT and not4D were treated with FLAG-tagged K48 or K63 linkage-specific Tandem Ubiquitin Binding Entities (TUBEs) before performing a-FLAG pull-down assays and Rps9a detection by a-GFP IB (54).
While only background levels of a high molecular weight (130-170 kDa) Rps9a-GFP specific signal was detected in the control and WT K48-TUBE specific IPs, the Rps9a-GFP signal increased specifically in the not4D K48-TUBE IP (Fig. 5B).Separately, the K63-TUBE IPs pulled down K63-linked ubiquitinated Rps9a above the background from WT, yet not4D did not alter its abundance (Fig. 5B).These data suggest that Not4 loss selectively increases K48-linked Rps9a ubiquitination.Importantly, the high molecular weight Rps9a-GFP signal is specific since the Rps9a-GFP signal was undetectable in the normal Rps9a-GFP size range of 56 kDa, thus indicating Rps9a is modified by multiple ubiquitin chains (Fig. 5B).While not4D results in incomplete Rps9a-GFP degradation, ccr4D results in mostly complete degradation and liberation of free GFP (Fig. 3C).How Not4 recognizes substrates is unknown, but a previous study determined that Ccr4 contributes to Not4-mediated ubiquitin signaling (55), suggesting that ccr4D likely impairs Not4-dependent inhibition of Rps9a-GFP degradation.To explore this possibility further, we tested if the Rps9a-GFP inhibition and its increased degradation also occurred in cells lacking the paralogous E2 enzymes Ubc4 and Ubc5 utilized by Not4 to mediate substrate ubiquitination (8).Rps9a-GFP expression was reduced similarly in both not4D and ubc4D ubc5D compared to WT cells, while ubc4D ubc5D caused even greater Rps9a-GFP partial degradation compared to not4D (Fig. 5C).Since Ubc4 and Ubc5 are not Ccr4-Not subunits, these data further support the concept that ccr4D impairs a Not4 ligase-regulated process that promotes 40S RP expression and inhibits 40S RP degradation.Since several different E3 ligases utilize these E2 enzymes (56), these results also indicate that additional E3 ligases must function redundantly with Not4 in this regard.
Because not4D increased Rps9a-GFP K48-linked ubiquitination, we tested if not4D inhibited Rps9a-GFP expression by increasing its proteasome-dependent degradation.WT and not4D were treated with 10 mM MG132 to inhibit the proteasome.Since WT cells are resistant to proteasome inhibitors, we also engineered a pdr5D into both WT and not4D and repeated these experiments.Pdr5 is a plasma membrane efflux pump that exports xenobiotics and H + from cells (57), and its loss sensitizes cells to proteosome inhibitors (58).While MG132 treatment did not inhibit proteasome activity in WT or not4D, pdr5D loss sensitized both WT and not4D to proteasome inhibition indicated by the MG132-dependent increase in global protein polyubiquitination (Fig. 5D).Unexpectedly, we found that not4D pdr5D restored Rps9a expression back to near WT levels irrespective of proteasome inhibition (Fig. 5, D and E).Longer a-GFP exposure also revealed that not4D pdr5D enhanced Rps9a-GFP degradation relative to not4D such that detectable free GFP (27 kDa) accumulated, which was not increased further by proteasome inhibition (Fig. 5D).These data reveal that not4D does not repress this 40S RP or promote its incomplete degradation through a proteasome-dependent mechanism.They also indicate that the combined pdr5D not4D opposes Rps9a repression due to not4D while it increases Rps9a-GFP degradation to free GFP.
Previously, we found that Ccr4-Not disruption inhibits TORC1 signaling by impairing vacuole V-ATPase activity and reducing vacuole acidity (30).V-ATPase-mediated vacuole acidification is essential for autophagy substrate degradation, nutrient storage and recycling, and the downstream activation of the RP and Ribi gene transcription and mRNA translation necessary for ribosomal biogenesis (59,60).Since Pdr5 exports H + from cells, we speculated that Pdr5 may oppose vacuole V-ATPase function indirectly by competing for a limiting pool of free cytoplasmic H + (61).If so, then in Ccr4-Not deficient cells where V-ATPase function is compromised (30), Pdr5 loss may alleviate the inhibitory effect that Ccr4-Not disruption has on the V-ATPase by reducing competition for H + .To test this possibility, not4D was combined with nhx1D that lacks the vacuole Na + -K + /H + exchanger, and causes vacuole hyperacidification (62).While not4D nhx1D reconstituted with NOT4 from a URA3 expression plasmid had no overt growth defects, selection on 5-FOA media, which forces NOT4 plasmid loss, revealed not4D nhx1D to be synthetically sick compared to not4D (Fig. 5F).However, the not4D nhx1D fully restored Rps9a-GFP expression, and it increased the conversion of the partial Rps9-GFP degradation product (40 kDa) to free GFP (27 kDa) (Fig. 5G).These results indicate that Ccr4-Not disruption represses RP expression and inhibits 40S RP degradation by disrupting V-ATPase-mediated acidification of the endolysosomal compartment.Furthermore, not4D ribosomal repression cannot be explained as an indirect consequence of poor growth since the not4D nhx1D restores Rps9a expression, yet it is more profoundly growth impaired (Fig. 5F).The Ccr4-Not ligase inhibits 40S ribosome autophagy degradation.Combined with the findings of increased partial (in not4D) and complete (in ccr4D) Rps9a-GFP degradation (Fig. 3C), we considered whether Ccr4-Not disruption represses RP levels through a macroautophagy-dependent mechanism.Initially, we tested if macroautophagy deregulation occurs in not4D by transforming WT, ccr4D, and not4D with a GFP-Atg8 expressing plasmid (32).GFP is cleaved from Atg8 when autophagosomes fuse with the vacuole, yet GFP resists vacuole degradation, so free GFP accumulation indicates increased macroautophagy (50).Log phase cultures of two independent plasmid transformants from each background were analyzed by a-GFP IB and, consistent with macroautophagy inhibition in nutrient-rich environments, WT cells have no detectable free GFP (Fig. 6A).As previously reported (32), ccr4D increased macroautophagy in nutrientrich conditions, which is indicated by the presence of free GFP (32), while not4D also activated macroautophagy (Fig. 6A).The ccr4D deregulates autophagy by stabilizing autophagy (ATG) transcripts, so ATG transcript levels were measured to determine if Not4 ligase signaling affects their expression similarly.ATG transcript levels were unaffected or elevated in both not4D + vector and not4D + NOT4RR cells, while their expression in not4D + NOT4 either was comparable to the WT + control or only slightly upregulated (Fig. S7).Thus, both Ccr4-Not ubiquitin ligase and deadenylase activities repress ATG transcript levels in macroautophagy-inhibitory environments.However, it remains unknown if Ccr4-Not ligase signaling accomplishes this by promoting their mRNA decay or through transcriptional repression.

40S ribophagy activation in
The deregulated macroautophagy in not4D and ccr4D suggested that the Rps9a-GFP degradation may be mediated by increased 40S ribophagy.To test this possibility, not4D and ccr4D were treated with the lysosomal inhibitor chloroquine which increases vacuole pH and inhibits autophagy (63,64).Chloroquine treatment reduced Rps9a-GFP degradation and prevented the liberation of free GFP in ccr4D, indicating that Ccr4-Not disruption mediates 40S degradation via vacuole-dependent autophagy (Fig. 6B).Intriguingly, chloroquine treatment did not alter the partial Rps9a-GFP degradation in not4D (Fig. 6B).To further probe a role for increased Rps9a-GFP autophagy in not4D, we treated both WT and not4D Rps9a-GFP and Srp40-GFP with the distinct autophagy inhibitor (PMSF) as previously reported (65).Consistent with the chloroquine results, PMSF treatment also failed to prevent the not4D dependent Rps9a-GFP partial degradation, while it completely prevented Srp40-GFP degradation to free GFP (Fig. 6, C and D).Ccr4-Not disruption decreases V-ATPase-mediated vacuole acidification (30), which is predicted to reduce autophagy efficiency due to the increased vacuole pH.We have observed that not4D causes greater V-ATPase inhibition relative to ccr4D (unpublished observations).Therefore, the above results would be consistent with not4D causing reduced autophagy efficiency of 40S RPs, most likely because they exist complexed with additional RPs, while single Ribi factors are degraded more efficiently.
To further explore the relationship between Ccr4-Not and 40S ribophagy, Rps9a-GFP WT, not4D, and ccr4D nutrientrich cultures were harvested, while a duplicate WT culture was treated with the TORC1 inhibitor rapamycin for 4 h to induce robust ribophagy.Although no Rps9a-GFP ribophagy occurred in the WT control, substantial free GFP (27 kDa) accumulated in both the rapamycin-treated WT and ccr4D cells (Fig. 6E).Importantly, an intermediate Rps9a-GFP degradation product was visible in the WT rapamycintreated sample, matching the size of the partial degradation product in not4D (Fig. 6E).These data provide further support for increased 40S ribophagy-mediated degradation of Rps9a-GFP in Ccr4-Not deficient cells.Since Ccr4-Not disruption inhibits TORC1 signaling (30), we tested if the 40S ribophagy was due solely to impaired TORC1 activity.WT and not4D were transformed with a control plasmid or plasmid expressing an HA-tagged Tor1 I1954V mutant that has increased kinase activity (66).Tor1 I1954V rescued WT and not4D growth in the presence of rapamycin (Fig. 6F), thus confirming it enhances TORC1 signaling in both.However, Tor1 I1954V failed to restore Rps9a-GFP expression or prevent 40S ribophagy in not4D (Fig. 6G).Therefore, while Ccr4-Not inhibits 40S ribophagy in nutrient-rich environments through a ubiquitin ligase-dependent mechanism, this 40S ribophagy inhibition is not due solely to Ccr4-Not mediated TORC1 activation.
Previously, not4D was reported to have a minor synthetic sick phenotype when combined with an atg7D mutant that blocks autophagosome-dependent macroautophagy (24).While an increase in 40S ribophagy was not reported in this study, a different 40S reporter strain (Rps3-GFP) was used in this analysis.We tested the possibility that the 40S ribophagy in not4D could be mediated by increased Atg7-dependent macroautophagy by combining not4D with atg7D.As previously reported, the atg7D not4D exhibited only a mild synthetic sick phenotype compared to not4D (Day 2, Fig. 7A).However, the not4D atg7D neither restored full-length Rps9a-GFP expression nor prevented Rps9a-GFP partial degradation (Fig. 7, B and C).Thus, while Ccr4-Not ligase disruption deregulates macroautophagy in nutrient-rich environments (Figs.3C, 4A, and 6), both the 40S RP repression and increased 40S ribophagy in not4D are Atg7 macroautophagyindependent.

TORC1-stress-induced 40S ribophagy requires a functional ESCRT pathway but redundant pathways mediate 40S ribophagy in a Ccr4-Not ligase mutant
The effectors selectively regulating 40S ribophagy remain unknown (51,67).Because the not4D vector and NOT4RR cells both induced partial 40S ribophagy, we analyzed their overlapping DEPs (Fig. 1D) for factors connected either to autophagy or ribosome turnover pathways that may mediate the 40S ribophagy.Three upregulated proteins were identified (Cue5, Nvj1, and Vps27) that have been linked to autophagy previously (Fig. S8).Cue5 is a CUET protein family member that binds ubiquitin and interacts with Atg8 to mediate proteaphagy and protein aggregate autophagy (68-70).Nvj1 is a The Ccr4-Not ligase inhibits 40S ribosome autophagy Figure 6.Ccr4-Not ligase deficiency causes 40S ribophagy independent of TORC1 inhibition.A, a plasmid expressing GFP-Atg8 from its native promoter was transformed into WT, not4D, and ccr4D.Log phase WCEs from two independent transformants for each strain background were analyzed by a-GFP IB.The GFP-Atg8 and GFP only bands are indicated.B, analysis of Rps9a-GFP in WT, not4D, and ccr4D mock treated or treated with 5 mM chloroquine (Chloro) for 4 h.WCE were prepared and 30 mg then analyzed by a-GFP IB.The a-G6PDH IB is the loading control.C, WT and not4D Rps9a-GFP cells were mock treated or treated (not4D) with 1 mM phenylmethanesulfonyl fluoride (PMSF) for 4 h before a-GFP IB as described in (A).D, as in (C) except Srp40-GFP was analyzed.E, the WT, not4D, and ccr4D strains expressing Rps9a-GFP were grown to log phase before preparing WCEs.In parallel, a log phase WT culture was treated with 200 nM rapamycin (Rap) for hours before harvesting and WCE preparation.Samples (30 mg) were analyzed by a-GFP IB and then a-G6PDH as a loading control.The black arrow indicates full-length Rps9a-GFP and the red arrows indicate partial or complete (27 kDa) degradation products.F, WT and not4D were transformed with control vector or vector expressing the Tor1 I1954V mutant.Equal numbers of cells from overnight cultures were 5-fold serially diluted and spotted to control selection plates or selection plates containing the TORC1 inhibitor rapamycin (10 nM).Images were acquired after 4 days of growth at 30 C. G, the strains from (F) were grown to log phase and WCEs (30 mg) then analyzed by a-GFP IB to detect Rps9a-GFP and a-HA to detect Tor1 I1954V expression.The a-G6PDH IB is the loading control.Short and long a-GFP IBs are indicated with the black arrow denoting full-length Rps9a-GFP and the red arrow indicating the partial degradation product.
The Ccr4-Not ligase inhibits 40S ribosome autophagy nuclear membrane protein that can interact with the vacuole protein Vac8 to promote piecemeal microautophagy of the nucleus (PMN) during nutrient stress (71).Vps27 is a subunit of the endosomal sorting complex required for transport (ESCRT) complex that binds ubiquitinated substrates to promote their degradation via autophagosome-independent microautophagy (72), while it also regulates endosomal protein sorting and turnover of plasma membrane proteins (73).Analysis by RT-qPCR revealed that loss of Not4 ligase activity increases the mRNA levels for all three proteins (Fig. 8A), indicating that Not4 ligase activity either promotes their mRNA decay or represses their transcription.To determine if they contribute to 40S ribophagy, WT, cue5D, nvj1D, and vps27D were treated with rapamycin for 4 h to induce ribophagy before analyzing Rps9a-GFP cleavage.While no Rps9a-GFP cleavage occurred in mock-treated cells, TORC1 inhibition induced robust Rps9a-GFP cleavage in WT, cue5D, and nvj1D, while this cleavage was abolished by vps27D (Fig. 8B).Thus, TORC1 inhibition activates 40S ribophagy dependent on a functional Vps27-dependent ESCRT pathway.60S ribosomes also undergo nutrient-stress induced ribophagy regulated by the Ltn1 ubiquitin ligase that ubiquitinates the 60S to inhibit 60S ribophagy, while the Ubp3 and Bre5 (Ubp3/ Bre5) deubiquitinase removes 60S ubiquitination to activate 60S ribophagy (67,74).Because 60S ribophagy was not detected in not4D or ccr4D (Fig. 3C), we inspected the Not4 proteome data to determine if the expression of these 60S ribophagy effectors was altered.Intriguingly, while Ltn1 and The Ccr4-Not ligase inhibits 40S ribosome autophagy The Ccr4-Not ligase inhibits 40S ribosome autophagy Ubp3 protein levels were unaffected, we found that Bre5 expression was repressed significantly, and it is among the 422 proteins that overlap in the not4D vector and NOT4RR cells (Figs.S9 and 1E).Overall, these data demonstrate that the ubiquitin-binding ESCRT subunit Vps27 is required for 40S ribophagy activation during nutrient stress, suggesting that the Vps27 upregulation in Ccr4-Not ligase mutants may contribute to 40S ribophagy.Our proteomic results also indicate that 60S ribophagy likely is not induced in Not4deficient cells due to Bre5 repression.
If Not4-deficient cells increase Vps27 expression as the sole mechanism for inducing 40S ribophagy, then combining vps27D with not4D should prevent its occurrence.To test this possibility, not4D was combined separately into vps27D, and also into cue5D or nvj1D for comparison, and then the growth phenotypes of the single and double mutants were compared.The cue5D not4D and nvj1D not4D all exhibited synthetic sick phenotypes when forced to lose the NOT4 expressing plasmid on 5-FOA plates, while the vps27D not4D was even more growth impaired (Fig. 8C).This extreme synthetic sick phenotype is consistent with a previous report that found mating vps27D to not4D failed to generate viable vps27D not4D spores (8).Surprisingly, when Rps9a-GFP expression was examined in the single and double mutants, each double mutant restored full-length Rps9a-GFP levels back to WT (Fig. 8, D and E).Furthermore, none of the double mutants, including vps27D not4D, prevented Rps9a-GFP ribophagy.Instead, each increased conversion of the partial Rps9a-GFP degradation product in not4D to free GFP (Fig. 8, D and E), indicating that loss of these factors partially alleviates the block in complete 40S ribophagy due to not4D.Therefore, although WT cells require Vps27 for 40S ribophagy during TORC1 stress, redundant mechanisms must activate this process in not4D.

Discussion
Cells coordinate anabolism and proliferation with their environment, which involves communication between the endolysosomal pathway, and the machinery controlling gene expression, translation, and proteostasis.Ccr4-Not regulates each of these processes, so it is uniquely positioned to integrate communication between these interrelated activities.While Ccr4-Not's role in transcriptome regulation is known, how Ccr4-Not ubiquitin signaling affects the proteome remains incompletely understood.To address this issue, we performed proteomic analysis of cells lacking Not4, expressing WT Not4, or expressing a full-length inactive Not4 mutant.This approach controls for proteome alterations indirectly caused by Ccr4-Not structural changes that may affect its non-ligase signaling roles.Indeed, we find that complete Not4 loss decreases expression of the Caf40 and Not2 subunits, thus indicating that Not4 loss likely destabilizes other Ccr4-Not subunits as previously reported for other Ccr4-Not mutants (43).In addition, while not4D + vector and not4D + NOT4RR cells share many DEPs (422 in total) that represent proteins dependent on Not4 ligase signaling, not4D + vector cells also have 300 DEPs that are not explained due to the loss of Not4 ligase activity.These differences likely are explained through indirect effects on non-ligase dependent Ccr4-Not regulated activities due to destabilization of other Ccr4-Not subunits in the absence of Not4.Alternatively, Not4 may contribute to mRNA transcription, stability, and/or mRNA translation independently of its role as a ubiquitin ligase.We also find that increased Not4 levels upregulate mitochondrial proteins, indicating that Not4 ligase signaling may mediate mitochondrial metabolic reprogramming, although how this occurs remains unclear.Since Ccr4-Not controls multiple aspects of mitochondrial regulation (30,39,40,75), future studies will need to address these important questions.
The two most intriguing findings presented herein are the RP and Rib repression due to dysregulated Ccr4-Not ligase activity, and the activation of 40S ribophagy when Ccr4-Not dependent ubiquitin signaling is lost.The repressed GOs for each experimental condition were dominated by ribosomal categories, clearly indicating that Not4 ligase signaling maintains ribosome homeostasis.These data are consistent with the previous report that Not4 loss inhibits polysome formation and alters ratios of 40S, 60S, and 80S ribosomes (24).Additionally, a previous metabolomic analysis found that the not4D metabolic profile is most similar to that of ribosomal mutants (76).These previous studies, and our data presented herein, provide increased support for the concept that Not4 ubiquitin signaling maintains ribosomal homeostasis.While Not4 overexpression restores many RP and Ribi factors, a core set remains repressed and overlaps those inhibited in both not4D + vector and not4D + NOT4RR cells (see Fig. S5 for visual representation).However, since Not4 overexpression corrects the protein synthesis defects and sensitivity to translational stress caused by not4D, their repression is insufficient to inhibit global protein synthesis.Another important observation is that the RP repression we detect is not selective solely for 40S or 60S RPs but it instead encompasses RPs from both.Our analysis of the model RP Rps9a indicates that neither proteasome nor macroautophagy inhibition restores Rps9a expression in not4D.Therefore, Not4 ligase activity does not inhibit RP expression by enhancing their degradation.Whether the mechanisms repressing both 40S and 60S RPs are the same or differ will need to be addressed in future studies, but it is clear this repression occurs post-transcriptionally.
We believe the most likely explanation for the RP and Ribi inhibitory effects is that Ccr4-Not ligase signaling affects RP and Ribi mRNA translation.Ccr4-Not binds to RP mRNAs during transcription elongation, which contributes to their cytoplasmic translation (43).If this process depends on Not4 ligase activity, then altering Not4 expression (too little or too much) and/or its function may impair RP mRNA translation.The RP inhibition detected also may be linked to the repression of specific Ribi factors that occurs since Ribi inhibition, perhaps due to defects in Ccr4-Not regulation of Pol I transcription (7), would be expected to cause ribosomal assembly defects (77).In this scenario, individual RPs and/or dysfunctionally assembled 40S ribosomes then could be flagged with ubiquitin to initiate their degradation via ribophagy.Altered The Ccr4-Not ligase inhibits 40S ribosome autophagy Not4 levels and ligase-dependent inhibitory effects on RP expression also is supported by our targeted analysis of Rps9a and Rpl36a since Not4 expressed from its chromosomal regulatory elements restores their expression to wild-type levels.Increased Not4 levels in wild-type cells also enhance sensitivity to select translational stressors that impair ribosome activity.These genetic data suggest the possibility that factors that interact with the ribosome to affect tRNA-ribosome interactions may be particularly sensitive to Not4 ligase signaling.In total, these data support the concept that balanced Not4 ligase activity maintains ribosomal homeostasis.Future efforts will need to deconvolute the underlying mechanisms to discern how Not4 ligase signaling mediates this control.
Another key finding in this report is that either Not4 or Ccr4 loss activates 40S ribophagy, indicating they repress 40S ribophagy through a shared mechanism requiring Ccr4-Not ubiquitin signaling.Ccr4 regulates endolysosomal nutrient signaling independently of its deadenylase activity (30), while it also contributes to Not4-dependent ubiquitin signaling (55).As such, Ccr4 loss impairs a key aspect of Not4 ubiquitin signaling that prevents 40S ribophagy activation.This possibility is further supported by the results demonstrating that loss of the E2 enzymes Ubc4/Ubc5, which are not Ccr4-Not complex members, causes even greater partial 40S ribophagy than not4D.These data further demonstrate that Ccr4-Not 40S ribophagy activation is due specifically to the disruption of its ubiquitin ligase activity, and they suggest that additional E3 ligases function redundantly with Not4 to inhibit this process.While Ccr4-Not disruption deregulates macroautophagy in nutrient-replete conditions, macroautophagy cannot explain the 40S ribophagy since atg7D not4D fails to prevent it.Furthermore, both Rps9a-GFP repression and 40S ribophagy in not4D are not due solely to TORC1 inhibition since restoring TORC1 signaling in not4D has no effect on either process.Previously, we found that Ccr4-Not activates TORC1 by promoting vacuole V-ATPase function (30).Our unpublished data indicate that not4D has a greater inhibitory effect on V-ATPase-dependent nutrient signaling than ccr4D, so additional V-ATPase nutrient signaling pathways promoting RP expression may be impaired specifically in not4D.Similarly, the incomplete 40S ribophagy in not4D likely a consequence of this greater V-ATPase inhibition since the vacuole-specific proteolytic activities required for efficient 40S ribophagy require an acidic vacuole.This possibility is supported by our finding that loss of the plasma membrane pump Pdr5, or the vacuole Na + -K + /H + exchanger Nhx1, both restores RP expression and enhances complete 40S ribophagy.These mutants likely compensate for the impaired V-ATPase activity in not4D by reducing competition for limiting intracellular H + or by decreasing vacuole H + efflux.Such an effect would decrease vacuole acidity, increase protein degradation, and promote the downstream vacuole-dependent nutrient signaling controlling RP expression.
Since the 40S ribophagy in not4D is independent of macroautophagy, it must result from the mechanistically distinct microautophagy pathway.K48 and K63 linked substrate polyubiquitination can signal degradation through both macroautophagy and microautophagy (22,23).Our data indicates that K48-linked ubiquitin chains accumulate on Rps9a in not4D, suggesting this ubiquitination may flag the 40S for ribophagy similar to that which occurs for proteaphagy.How K48-linked 40S ubiquitination accumulates is not known, but it may be connected to the defects in proteasome-dependent deubiquitination that occur in Not4 mutants.Ribosome downregulation mediates adaptation to proteostatic stress (78), so the combined RP/Ribi repression and 40S ribophagy activation in not4D may allow adaptation to the increased proteostatic stress due to not4D proteosome dysregulation.
Three proteins previously connected to various autophagy pathways (Cue5, Nvj1, and Vps27) are upregulated in not4D vector and NOT4RR cells where 40S ribophagy is activated.Importantly, only loss of Vps27, which is an ESCRT ubiquitin binding subunit that is both required for vesicular transport and also is a microautophagy effector, blocked TORC1 stressinduced 40S ribophagy in wild-type cells.To our knowledge, this is the first factor identified to be required for 40S ribophagy.However, since 40S ribophagy was not prevented by the vps27D not4D mutant, redundant pathways must mediate 40S ribophagy in not4D.Surprisingly, the 40S RP degradation defect in not4D was partially restored by vps27D not4D, cue5D not4D, and nvj1D not4D.How co-disruption of these pathways enhances 40S RP degradation currently is not understood, but their effect appears similar to that which occurs upon loss of Pdr5 or Nhx1.A common theme linking them is their functional connection to endolysosomal-dependent nutrient signaling and stress pathways.Cue5 is a ubiquitin-binding receptor for proteaphagy and aggrephagy (68,69), yet it has extensive genetic interactions with endolysosomal factors indicating it may have endolysosomal functions not yet identified (37,79).Cue5 loss in not4D may alter these other activities to restore RP expression and enhance 40S ribophagy.Although Nvj1 mediates PMN, it also interacts with the oxysterol-binding protein Osh1 which promotes nutrient transport from the plasma membrane (71).A nvj1D enhances nutrient uptake by preventing sequestration of Osh1 from this additional role (80), so the nvj1D not4D may enhance nutrient signaling to compensate for the not4D defects.Since Vps27 also mediates vesicular transport, the vps27D may increase nutrient transporters on the cell surface as their turnover could be reduced through the endolysosomal pathway (18), which also would enhance endolysosomal nutrient signaling.Thus, the vacuole defects in not4D may be alleviated by the loss of these factors through multiple candidate mechanisms.
During the course of this study, a SILAC-based analysis of the WT and not4D proteome found that not4D increases the expression of several RPs relative to WT cells (42), which is a result seemingly at odds with our proteomic data.Previous studies using specialized protein extraction approaches found that not4D can form protein aggregates in vivo, which may be related to the loss of Rps7a monoubiquitination (24,25).However, to prevent aggregates we prepared the proteomic samples by bead beading in a high salt buffer.We also directly demonstrate that both Rps9a and Rpl36a are repressed in The Ccr4-Not ligase inhibits 40S ribosome autophagy not4D even when analyzed under denaturing conditions.Additionally, ribosomal protein aggregation would not account for the significant RP and Ribi repression in not4D + NOT4 reconstituted cells where Rps7a monoubiquitination is intact, nor would it explain why disrupting endolysosomal effectors in not4D restores RP expression.Future efforts will be required to reconcile the differences between our results and this recent report.
Analysis of Ccr4-Not has focused on its role in transcription and mRNA decay.However, increasing evidence indicates that its ubiquitin signaling role is critically important to additional cellular processes.The data presented herein demonstrate that balanced Not4 ligase activity is required to maintain RP expression, while the Ccr4-Not ubiquitin ligase inhibits 40S ribophagy.Combined with our previous study connecting Ccr4-Not to endolysosomal regulation, these new data reveal increased importance for the Ccr4-Not ligase in endolysosomal-dependent nutrient signaling and autophagy control than currently is appreciated.Since Ccr4-Not dysregulation is implicated in many diseases and developmental disorders, probing these additional Ccr4-Not regulated processes will be essential to understanding Ccr4-Not's role in these conditions.

Yeast strains and culture conditions
All strains in this study are derived from the BY4741 or DF5 genetic backgrounds and are listed in Table S1 (81).For all experiments, cells were cultured in an incubated shaker at 30 C and harvested between OD 600 = 0.8 to 1.6.Cells were cultured in nutrient-rich YPD (1% yeast extract/2% peptone/ 2% dextrose).To select for plasmid transformants, cells were cultured in synthetic complete (SC) media containing 0.2% yeast nitrogen base without amino acids, 0.5% ammonium sulfate, 2% dextrose, and 0.19% amino acid dropout mix with amino acids or uracil added back except the specific one needed to maintain plasmid selection.For the puromycin labeling experiment, puromycin was added to the cultures at a final concentration of 10 mM and samples were taken at the indicated timepoints.The genetic manipulations used to either delete target genes or to engineer genomically integrated inframe epitope tags, were performed as described (82).All yeast plasmids used in this study are listed in Table S2.

Chemical reagents
Most chemicals, including rapamycin, Hygromycin B, and puromycin, were purchased either through Fisher or Millipore Sigma.Yeast media was purchased from US Biologicals and 5-Fluoroorotic acid (5-FOA) was purchased from Research Products International.

RNA extraction and RT-qPCR analysis
Total RNA was prepared by hot acid phenol extraction from three independent biological replicates per experimental condition.The ImProm-II reverse transcription kit (Promega) was used to synthesize cDNA from 1 mg of total RNA using the included oligo dT primers in a volume of 20 ml per the manufacturer's direction.After cDNA synthesis, the cDNA reaction was diluted with ultrapure dH 2 O to a final volume of 100 ml, and 1 ml of the diluted cDNA was used to perform gene-specific qPCR using SYBR Green reagent and an ABI StepOne Plus thermal cycler.The gene-specific signal was normalized to the SPT15 internal reference control gene using the formula 2 (CtSPT15-CtTarget) as previously described (83).The mean and standard deviation for each gene were calculated and are plotted in Figs.S3 and S7.Primer sequences used in this study are in Supplemental File 1.All statistical analyses and graphs were generated using GraphPad Prism 10.

Antibodies and immunoblotting reagents
All immunoblots are representative of at least three independent experiments.The a-HA (clone F-7, catalog number sc7392), a-GFP (clone B-2, catalog number sc9996), and aubiquitin (clone P4D1, catalog number sc8017) antibodies were purchased from Santa Cruz Biotechnology.The a-Rps3 antibody was purchased from Proteintech (catalog number 11990-1-AP), while the a-FLAG (clone M2, catalog number F1804) and a-glucose-6-phosphate dehydrogenase (G6PDH, catalog number A9521) antibodies were purchased from Millipore Sigma.The a-puromycin antibody (clone PMY-2A4) was purchased from the Developmental Hybridoma Studies bank.Mouse and rabbit HRP-conjugated light chain-specific secondary antibodies were purchased from Jackson Immu-noResearch Laboratories (catalog numbers 115-035-174 and 211-032-171).Immunoblots were performed using Milli-poreSigma Immobilon Western Chemiluminescent HRP Substrate and images were quantified using ImageJ.

Whole-cell extract preparation
Whole-cell extracts (WCEs) were prepared by bead beating in extraction buffer (300 mM NaCl, 10 mM Tris, pH 8.0, 0.1% NP-40, 10% glycerol) and containing protease inhibitors (2 mg/ ml each of aprotinin, pepstatin, and leupeptin), phosphatase inhibitor cocktail set II (Millipore Sigma, 1:100 dilution) and 1 mM DTT. Cell lysates were clarified for 15 0 at 15,000 rpm @ 4 C, supernatants then were transferred to a new tube, and the protein concentration was determined by Bradford assay.

TUBE ubiquitin linkage analysis
WT and not4D Rps9a-GFP expressing cells were cultured to log phase in YPD media, harvested, and WCEs then prepared by bead beating in extraction buffer (150 mM NaCl, 10 mM Tris, pH 8.0, 0.1% NP-40, 10% glycerol, and 60 mM 2-chloroacetamide) containing protease inhibitors (2 mg/ml each of aprotinin, pepstatin, and leupeptin), and phosphatase inhibitor cocktail set II (Millipore Sigma, 1:100 dilution).For each sample, 600 mg of WCE was mixed with 2 mg of a-FLAG antibody in a total volume of 400 ml WCE buffer that had been mock treated or treated with 3 mg of K48ub or K63ub linkage specific Tandem Ubiquitin Binding Entities (TUBEs) purchased from LifeSensors.Samples were rotated at 4 C for a minimum of 2 h before adding Protein A agarose beads and The Ccr4-Not ligase inhibits 40S ribosome autophagy rotating at 4 C for an additional 1 h to capture the immune complexes.Samples were washed three times with buffer and then resolved by 10% SDS-PAGE and analyzed by a-GFP immunoblotting.

Confocal microscopy
The procedure for fluorescence microscopy was adopted from methods previously described with slight modifications (84,85).The Rps9a-GFP wild-type strain was grown in 5 ml SC media overnight with shaking at 30 C. In the morning, 1 ml SC media was inoculated such that the initial OD 600 was approximately 0.1.Immediately after inoculation, 1 ml FM4-64 dye (from 1 mg/ml stock in DMSO from Invitrogen) was added to the culture, and tubes were incubated as above.After 5 to 6 h, cells were pelleted by centrifuging at 5000g for 5 min, washed twice with PBS pH7.4,and the cell pellet then was resuspended in a small volume of PBS pH7.4.A 5 ml cell suspension was placed on a glass slide, mixed with 5 ml of 1% low melting agarose (maintained in a 37 C water bath), covered with a cover slip, and images were captured immediately using a Zeiss LSM 700 confocal microscope with a 63× objective.Image analysis was performed using Zeiss Zen Blue software.

Tandem mass tag labeling and quantitative proteomic analysis
WCEs were prepared in 300 mM NaCl lysis buffer as described above.Samples (50ul, 50ug protein) then were reduced with 5 mM DTT for 45 min at 50 C, alkylated with 25 mM iodoacetamide for 20 0 at room temperature in the dark, and then incubated with 20 mM DTT for an additional 15 0 at room temperature.Samples then were precipitated with five volumes of cold acetone (at −20 C overnight), and proteins were pelleted at 16,000g at 4 C for 10 0 .Protein pellets were washed with 100 ml of cold (−20 C) 90% acetone, air dried for 4 minutes, and re-dissolved in 100 ml of digestion buffer (100 mM HEPES, pH 8.3) containing 0.5 mg of Lys-C enzyme (Waco, Fuji, 125-05061).The protein mix then was digested for 2 h at 37 C with shaking; adding 1ug trypsin (Thermo Fisher, 900057), the digestion was continued overnight.
For TMTpro-16plex labeling, a set of 16 samples each containing 50 mg of peptides in 100 ml of digestion buffer (100 mm HEPES pH 8.3) was labeled using a commercial TMTpro-16plex Mass Tag Labeling reagent kit (A44521, Thermo Fisher) according to the manufacturer's protocol.The labeled set of 16 samples was combined, vacuum-dried, and reconstituted in 0.1% TFA at 0.3 mg/ml for further fractionation.300 ml (90 mg) of reconstituted mixture of labeled peptides was fractionated using Pierce High pH Reversed-Phase Peptide Fractionation kit (84,868, Thermo Fisher) according to the manufacturer's protocol for TMT-labeled peptides, with eight-step fractions (consecutively eluted with 10.0, 12.5, 15.0, 17.5, 20.0, 22.5, 25.0, and 50.0%acetonitrile) collected.The peptide fractions were vacuum dried, dissolved in 65 ml of loading buffer (3% ACN with 0.1% TFA), and 5 ml (0.87 mg, assuming elution of equal amounts of peptides per fraction) aliquots were analyzed by LC-MS for peptide/protein identification and quantification.Acquisition of raw MS data was performed on an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher) operating in line with Ultimate 3000RSLCnano UHPLS system (Thermo Fisher) using MS2 method for TMTpro-16plex labeled samples with 160 min LC gradient.The peptides were trapped on an Acclaim PepMap 100 nanoViper column (75 mm × 20 mm, Thermo Fisher) at 5 ml/min flow rate and washed with loading buffer for 5 min.The trapped peptides were separated on an Acclaim PepMap RSLC nanoViper column (75 mm × 500 mm, C-18, 2 mm, 100 Å, Thermo Fisher) at 300 nl/min flow rate and 40 C column temperature using water and acetonitrile with 0.1% formic acid as solvents A and B, respectively.The following multi-point linear gradient was applied: 3% B at 0-4 min, 5% B at 5 0 , 23% B at 110 0 , 30% B at 120 0 , 90% B at 123 to 133 0 , 3% B at 136 to 160 0 .MS2 acquisition method was used with three-second cycles with the following MS scan parameters.Full MS scans were performed in the Orbitrap analyzer at 120,000 (FWHM, at m/z = 200) resolving power to determine the accurate masses (m/z) of peptides.Following MS1 full scans, data-dependent MS2 scans were performed on precursor ions with peptide isotopic pattern, charge state 2 to 6, and intensity of at least 25,000.Peptide ions were isolated in the quadrupole with 0.7 m/z window, fragmented (HCD, 38% NCE), and the masses/intensities of fragment and reporter ions were determined with an Orbitrap analyzer at 50,000 (FWHM, at m/z = 200) resolving power and dynamic exclusion applied for 45 s.
Post-acquisition analysis of raw MS data was performed within a mass informatics platform Proteome Discoverer 2.4 (Thermo Fisher) using Sequest HT search algorithm and yeast protein database (SwissProt, Saccharomyces cerevisiae S288c, TaxID 559292, v.2017-10-25, 6727 entries).The reversed target database was used as decoy database.The raw MS data acquired for the set of eight fractions (derived from the same mixture of labeled samples) were treated as 'Fractions' for post-acquisition analysis.Full tryptic peptides were searched and two mis-cleavages were allowed.The searched fixed modifications included: carbamidomethylation of Cys residues and TMTpro modification of Lys residues and any peptide Nterminus.The variable modifications included oxidation of Met residues, acetylation, and Met-loss of the protein N-terminus.The precursor, fragment, and reporter ion mass tolerances were set to 10 ppm, 0.02 Da, and 20 ppm, respectively.The raw data were filtered for the precursor ions with S/N of at least 1.5.The PSMs were filtered for further analysis using a delta Cn threshold of 0.05.The q-values were calculated at PSM level (Percolator) and then, at the peptide level (Qvality algorithm), to control for the false discovery rate (FDR).The FDR threshold of 0.01 was used to validate and filter the PSMs and the corresponding peptide sequences.The validated/ filtered peptides were used for the identification of the candidate precursor proteins.Each candidate precursor protein was scored by summing the PEP values of the assigned peptides.The sum-PEP protein scores were used to calculate the experimental q-values at the protein level.The candidate The Ccr4-Not ligase inhibits 40S ribosome autophagy proteins were further validated using 0.01 (strict) and 0.05 (relaxed) FDR thresholds.A set of candidate proteins was accepted as an identified protein group if none of the assigned peptides was unique to any protein, but at least one of those peptides was common and unique to that group of proteins.Strict parsimony principle was applied to the protein groups.The reporter ions quantification values (based on S/N ratio) were corrected according to the product data sheet provided with the used TMTpro-16plex Label reagent set.The coisolation and average S/N ratio thresholds were set to 50% and 10, respectively.Unique and razor peptides were used for protein quantification, and protein groups were considered for peptide uniqueness.

Proteome data analysis
The data generated by the mass spectrometry analysis was transferred to the UTHSC Molecule Bioinformatics Core using SFTP.These data then were normalized using the normal-izeCylclicLoess function from R/Bioconductor-package limma (86).The normalized data matrix was loaded into R, and statistics were gathered and used to determine differential expression.The mean, variance, standard deviation, and standard error of means were calculated for each protein across the different conditions.Pearson's correlation coefficient was graphed in order to identify sample outliers in each condition.At this time, no outliers were determined.A principal component analysis was performed to determine different clusters in the samples, and fold change was then calculated for all proteins.The significance for each protein was assessed by using one-way ANOVA, and the p values were adjusted for multiplicity using the Benjamini-Hochberg method (87).Only proteins with an adjusted p value < 0.05 were considered differentially expressed (File S2).These proteins are used to create a heatmap with unsupervised hierarchical clustering.The protein list then was loaded into the STRING database for protein-protein interaction mapping and to test for enriched cellular component GO categories (44).

Figure 1 .
Figure 1.Not4 ligase levels and activity differentially regulate the cellular proteome.A, wild-type (WT) cells transformed with control vector, and not4D cells expressing either control, NOT4, or NOT4RR vector, were streaked to duplicate plates and incubated at control temperature (30 C) or heat shock temperature (37 C) for 3 days before image capture.B, reverse transcriptase coupled with quantitative PCR (RT-qPCR) analysis of NOT4 transcript levels from the indicated strains.The NOT4 specific signal was normalized to the expression of the internal control gene SPT15 and significance was tested by oneway ANOVA.*-p< 0.05 or greater.C, immunoblot of 30 mg whole cell extract from the cells used in B. Arrow indicates the Not4 band and the asterisk indicates a non-specific band.D, proteome principal component analysis (PCA) of the individual replicates for the indicated strains.Each experimental condition had four independent biological replicates except for not4D + NOT4, which had three independent biological replicates.E, the differentially expressed proteins (DEPs) were identified by using a 1.5-fold change in expression and an FDR < 0.05.The resulting DEPs then were subjected to Venn analysis.The total number of DEPs in each condition relative to the WT + vector control is 959 (not4D + vector), 290 (not4D + NOT4), and 701 (not4D + NOT4RR).

Figure 2 .
Figure 2. Ribosomal protein (RP) and ribosomal biogenesis (Ribi) factors are repressed by changes to Not4 activity or expression levels.A-C, bubble blot analysis of the top cellular component gene ontology (GO) categories for the indicated experimental conditions.GO analysis was performed through the STRING database.File S3 has the GO category and individual protein results used to generate these bubble blots.D, Venn analysis of the reduced RPs identified in this study versus the RPs differentially expressed in not4D from the Allen study(42).E, heatmap analysis of the Log 2 expression for each overlapping RP identified in (D).File S3 has the RPs listed used for this analysis.

Figure 3 .
Figure 3.Both Not4 ligase activity and expression level inhibit RPs and affect ribosomal function in vivo.A, the indicated cells were grown to log phase and labelled with 10 mM puromycin before taking samples at the indicated timepoints.Whole cell extracts (WCEs, 30 mg) then were analyzed by apuromycin immunoblot (IB).The a-G6PDH IB serves as a loading control.B, WT and not4D transformed with the indicated expression vectors were cultured overnight at 30 C. Equal numbers of cells then were 5-fold serially diluted and spotted onto the indicated plates.The number of days the plates were incubated at either 30 C or 37 C before pictures were taken are indicated in parentheses.C, WCEs (30 mg) from the indicated strains in the Rps9a-GFP or Rpl36-GFP reporter backgrounds were analyzed by a-GFP and a-G6PDH IB.D, as in (B) except the indicated strains were spotted to control plates or plates containing 70 mg/ml Hygromycin B. E-G, As in (C) except Rps30b-GFP and Rps3 (E), Srp40-GFP (F), or Prp11-GFP (G) were analyzed.

Figure 4 .
Figure 4. Not4 expression level and ligase activity differentially affect 40S RP expression and degradation.A, duplicate IBs of WCEs (30 mg) from the indicated strains were probed with a-GFP to detect Rps9-GFP and a-FLAG to detect Not4 or Not4RR expression.The a-G6PDH IBs serve as the loading control for each.Short and long exposures for the a-GFP IB are presented.Black arrow indicates full-length Rps9a-GFP and red arrow indicates partially degraded Rps9-GFP.The * in the a-FLAG IB indicates a non-specific band.B, the mean and SD of the peptide abundance signal from the proteomic data for Rps9a of the individual replicates.****p < 0.001 by one-way ANOVA.The Log 2 fold-change (FC) for the individual not4D samples relative to the WT + vector are as follows: +vector (−0.87); +NOT4 (−1.6); +NOT4RR (−1.2).C, same as in (A) except Rpl36a-GFP is analyzed.D, the mean and SD of the peptide abundance signal from the proteomic data for Rpl36a of the individual replicates.****p < 0.001 by one-way ANOVA.The Log 2 fold-change FC for the individual not4D samples relative to the WT + vector are as follows: +vector (−1.34); +NOT4 (−1.8); +NOT4RR (−1.34).

Figure 5 .
Figure 5. Not4 ligase loss causes Rps9a degradation independently of the proteasome but dependent on the endolysosomal pathway.A, equalnumbers of WT and not4D were lysed by boiling in SDS-PAGE loading buffer and then analyzed by a-GFP IB.The a-G6PDH IB is the loading control.B, all WCEs (600 mg) from the indicated strains were treated with a-FLAG antibody combined in the absence or presence of 3 mg of either K48 or K63-specific TUBES.After incubation with rotation at 4 C for 2 h, protein A conjugated agarose beads were added and incubated for an additional 1 h.The beadimmune complexes then were pelleted and washed extensively before analyzing by 10% SDS-PAGE and a-GFP IB.The inputs represent 30 mg of the WCEs.Short and long a-GFP exposures are provided.The a-G6PDH IB serves as the loading control.*-non-specific band.C, log phase WCEs (30 mg) were prepared from the WT and indicated single or double mutants and analyzed by a-GFP.D, the indicated strains were grown to log phase and then mock treated or treated with 10 mM MG132 for 2 h to inhibit the proteasome.Short and long a-GFP IBs are provided.Black arrow indicates full-length Rps9a-GFP and red arrows indicate partial and complete degradation products.The a-G6PDH IB is the loading control.E, quantitation of (D).The short a-GFP IBs for a total of 3 to 4 individual biological replicates were quantified and expressed as the ratio of full-length GFP to G6PDH.*p < 0.05 by one-way ANOVA.F, the WT and not4D strains were transformed with control or NOT4 expression vector.Equal numbers of cells from overnight cultures then were 5-fold serially Ccr4-Not ligase-deficient cells is independent of TORC1 inhibition and the Atg7-dependent macroautophagy pathway Our results indicate that the increased K48-linked RP ubiquitination in not4D does not result in RP proteasomal diluted and spotted to either plasmid-selective media or media containing 5-FOA and images were captured on the indicated days.G, log phase WCEs (30 mg) from the indicated WT, single, and double mutants were IB with a-GFP.Short and long exposures are provided with the black arrow denoting fulllength Rps9a-GFP and the red arrows indicating degradation products.The a-G6PDH IB is the loading control.

Figure 7 .
Figure 7. 40S ribophagy activation in Ccr4-Not ligase deficient cells is independent of macroautophagy.A, the individual and double mutants were cultured overnight, and then equal cell numbers were 5-fold serially diluted, spotted to YPD plates, and incubated at 30 C. Images were taken at the indicated times.B, Log phase WCEs (30 mg) from the strains in (A) were analyzed by a-GFP IB.Short and long exposures are provided with the full-length Rps9a-GFP indicated by the black arrow and the degradation product denoted by the red arrow.C, quantification of (B).The short a-GFP IBs for a total of 3 to 4 individual biological replicates were quantified and expressed as the ratio of full-length GFP to G6PDH.**p< 0.005 or greater by one-way ANOVA.

Figure 8 .
Figure 8.Both 40S RP repression and ribophagy activation in Ccr4-Not ligase deficient cells are functionally connected to the endolysosomal pathway.A, RT-qPCR analysis of the indicated genes from triplicate independent biological replicates with the mean and SD plotted and significance determined by one-way ANOVA.**p < 0.005 or greater.B, the WT and individual mutants were cultured to log phase and mock-treated or treated with 200 nM rapamycin for 4 h before harvesting and performing a-GFP IB.The black arrow indicates full-length Rps9a-GFP and the red arrow indicates free GFP.C, WT cells and the indicated double mutants were transformed with either control or NOT4 expression vector and were cultured overnight.Equal cell numbers then were 5-fold serially diluted, spotted to plasmid-selective control media or to media containing 5-FOA, and then incubated at 30 C for the indicated times before photographing.D and E, the WT, single, and double mutants were grown to log phase and then analyzed by a-GFP IB.Short and long a-GFP exposures are provided.The black arrow indicates full-length Rps9a-GFP and the red arrows indicate partial and complete GFP degradation products.