Aggregation of CAT tails blocks their degradation and causes proteotoxicity in S. cerevisiae

The Ribosome-associated Quality Control (RQC) pathway co-translationally marks incomplete polypeptides from stalled translation with two signals that trigger their proteasome-mediated degradation. The E3 ligase Ltn1 adds ubiquitin and Rqc2 directs the large ribosomal subunit to append carboxy-terminal alanine and threonine residues (CAT tails). When excessive amounts of incomplete polypeptides evade Ltn1, CAT-tailed proteins accumulate and can self-associate into aggregates. CAT tail aggregation has been hypothesized to either protect cells by sequestering potentially toxic incomplete polypeptides or harm cells by disrupting protein homeostasis. To distinguish between these possibilities, we modulated CAT tail aggregation in Saccharomyces cerevisiae with genetic and chemical tools to analyze CAT tails in aggregated and un-aggregated states. We found that enhancing CAT tail aggregation induces proteotoxic stress and antagonizes degradation of CAT-tailed proteins, while inhibiting aggregation reverses these effects. Our findings suggest that CAT tail aggregation harms RQC-compromised cells and that preventing aggregation can mitigate this toxicity.

A conserved back-up degradation pathway mediated by Rqc2 and its prokaryotic homologs mitigates some of the toxicity associated with loss of tmRNA or Ltn1 function [24,35]. Rqc2 homologs bind the large ribosomal subunit and direct it to elongate the incomplete polypeptide's C-terminus with either alanine ("Ala tails" in prokaryotes) or both alanine and threonine residues ("CAT tails" in yeast) [24,36,37]. Metazoan CAT tails may include a more diverse repertoire of amino acids [34]. These extensions, made without a small subunit or mRNA, act as "degrons" to mark incomplete polypeptides for degradation by the bacterial protease ClpXP or the eukaryotic proteasome [24,35].
Loss of Ltn1 function results in a buildup of CAT-tailed ("CATylated") proteins [36]. The CAT tails on these proteins can homo-typically associate and form aggregates of CATylated proteins, which have been observed in yeast and metazoans [26,34,38,39]. The presence of threonine in yeast CAT tails increases their aggregation propensity [38]. It is thereby an attractive hypothesis that CAT tail aggregation has an adaptive function during stress. For instance, aggregation may protect the proteome by sequestering potentially harmful undegraded RQC substrates when Ltn1 becomes overburdened. Alternatively, it has been hypothesized that CAT tail aggregates disrupt cellular fitness by depleting the pool of chaperones [26,38,39], and thereby contribute to the proteotoxicity associated with Ltn1 failure. No study has evaluated these hypotheses by controlling the aggregation state of CAT tails and testing whether CAT tail aggregation plays a beneficial or harmful role during stress.
We used chemical and genetic approaches in Saccharomyces cerevisiae to modulate CAT tail aggregation propensity in Ltn1-deficient cells and determine how aggregation affects: 1) the function of CAT tails as degrons and 2) proteotoxic stress. We found that under conditions favoring CAT tail aggregation (elevated Rqc2 levels), CAT tails lose their degron function and induce proteotoxic stress. Conversely, we observed that increasing CAT tail solubility (with guanidinium hydrochloride or by genetic perturbation of RNA Polymerase III function) restores CAT tail-mediated degradation and reduces proteotoxic stress. Our work demonstrates that CAT tail aggregation contributes to the toxicity of Ltn1 disruption rather than aiding in adaptation to stress.

Plasmids
Plasmids were constructed using NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs, Ipswich, MA). All plasmids used in this study are described in S2 Table. Two variants of RQCsub were used that had identical construction except for the region following the stall: one had a fluorescent protein (RFP) and the other had a series of affinity tags. The version without the post-stall RFP was used for microscopy and the version with the poststall RFP was used in immunoblots.

Flow cytometry
Fluorescence was measured using a BD Accuri C6 flow cytometer (BD Biosciences). Three independent yeast cultures were grown to log phase overnight in appropriate SD dropout media. Data analysis was performed with MATLAB (MathWorks, Natick, MA).
For measurements of the Hsf1 activity reporter, raw fluorescence values were first normalized to side scatter, log 2 -transformed, and then the value obtained from the wt strain was subtracted from each condition. These calculations are reported as log 2 fold wt in the figures.
For measurements of RQCsub LONG , a detailed description of the analysis workflow can be found in [35]. Briefly, yeast expressing the plasmid were gated according to above-background RFP fluorescence. Then, bleedthrough from the RFP channel was calculated using an RFPonly control strain and subtracted from signal in the GFP channel. The resultant bleedthrough-corrected GFP:RFP values were normalized to the corresponding untreated hul5Δ strain for each background. To quantify the change in GFP:RFP upon HUL5 deletion, the normalized value from the HUL5 condition was subtracted from that of the hul5Δ condition.
For the stalling reporter, the analysis workflow matched that of RQCsub LONG with the order of the fluorophores being reversed, i.e. plasmid-expressing yeast selected based on GFP, bleedthrough from GFP subtracted from RFP. Where indicated, corrected RFP:GFP values were normalized to a non-stalling reporter that contained a non-stalling sequence encoding serine and threonine [17] in place of the stalling CGN codons.

Statistical analysis
A paired t-test (using the "ttest" function in MATLAB) was used to assess whether mean measurements differed between two different conditions, using the null hypothesis: μ condition 1 = μ condition 2 . To calculate the s.e.m. for Δ stability measurements, a propagation of error formula was used: SEM hul5Δ-HUL5 = sqrt(SEM HUL5 2 + SEM hul5Δ 2 ). To analyze the significance of Δ stability measurements, a t-test for particular contrast was performed using the "lm" and "linearHypothesis" functions in R (R Foundation for Statistical Computing). The null hypothesis for this test was: μ hul5Δ+ condition 1 -μ HUL5 + condition 1 = μ hul5Δ+ condition 2 -μ HUL5 + condition 2 .
For SDS-PAGE, samples were loaded into a NuPAGE Novex 4-12% Bis-Tris 1.5 mm protein gel (Thermo Fisher Scientific) and run in MOPS buffer. Gels were semi-dry transferred to 0.45 μm nitrocellulose membranes (Thermo Fisher Scientific) using the Trans-Blot Turbo system (Bio-Rad, Hercules, CA). Gels were wet transferred to .45 μm nitrocellulose membranes (Thermo Fisher Scientific) in the Trans-Blot Cell (Bio-Rad) in Towbin Buffer or semi-dry transferred using a Trans-Blot Turbo (Bio-Rad).

Lysate preparation and immunoprecipitation (IP)
Yeast were grown in SD media and harvested at OD600 0.8 to 1.0. Cells were harvested using either 1) centrifugation followed by resuspension in IP buffer (100 mM KOAc, 10 mM MgCl2, 25 mM HEPES-KOH pH 7.4) and freezing of cell droplets in liquid nitrogen; or 2) vacuum filtration and flash freezing in liquid nitrogen. Frozen yeast were cryogenically pulverized into powder using a Freezer/Mill (SPEX SamplePrep, Metuchen, NJ) and stored at -80˚C.
Frozen yeast powder was thawed and solubilized at 4˚C in IP buffer supplemented with Pierce Protease Inhibitor Tablets, EDTA-free (Thermo Fisher Scientific). Crude lysate was clarified by a single spin at 5000 x g. For lysate immunoblots, this lysate was denatured by addition of 4X NuPAGE LDS Sample Buffer with 5% β-mercaptoethanol and boiling, then analyzed as described above in "Immunoblots." For immunoprecipitations, the lysate was incubated with GFP-Trap_A (ChromoTek, Planegg-Martinsried, Germany) resin for 1 hour at 4˚C with rotation, then washed 10 times in IP buffer. For immunoblotting, washed GFP-Trap resin was boiled in 4X NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) with 5% βmercaptoethanol and analyzed as described above in "Immunoblots."

TEV protease digestion
GFP IP was performed as described above, but instead of immediate elution, the resin was equilibrated in IP buffer with 1 mM DTT and incubated overnight at 4˚C with 5 μL ProTEV Plus (Promega, Sunnyvale, CA) in 100 μL total reaction volume. The resin was then washed with fresh IP buffer with 1 mM DTT and boiled in 4X NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) with 5% β-mercaptoethanol and analyzed as described above.

Fluorescence microscopy
Live cell fluorescence imaging was performed using two set-ups: 1) a Nikon Eclipse Ti-E inverted fluorescence microscope (Nikon, Melville, NY) using a 100X/1.4NA oil objective lens and μManager [40], and 2) a Nikon Eclipse 80i microscope using a 100X/1.4NA oil objective lens and Metamorph (Molecular Devices, San Jose, CA). Yeast were grown in SD media to log phase, centrifuged briefly, resuspended in 5 μL fresh SD media and mounted on a clean coverslip and slide. Images were prepared and analyzed with ImageJ (NIH). For quantification, images were randomly acquired until N > 300 cells. Independent cultures were imaged on different days and error represents standard deviation of 3 independent cultures.

Yeast spot assay
Yeast were grown to log-phase in YPD, then diluted to OD600 = 0.1. 200 μL of this diluted culture was diluted 1:10 into YPD four times in a sterile 96-well plate (Greiner Bio-one, Kremsmünster, Austria). 5 μL of the serial dilutions were transferred from the 96-well plate onto YPD agar plates with or without 5 mM guanidinium hydrochloride (Sigma-Aldrich), 50 ng/ mL cycloheximide (Sigma-Aldrich) (with and without additional 5 mM guanidinium hydrochloride), 1.5 mM paromomycin (Chem-Impex International, Wood Dale, IL), 50 μg/ml hygromycin B (Goldbio, St. Louis, MO). Plates were then incubated at 30˚C or 37˚C. Plates grown at 30˚C with or without 5 mM guanidinium hydrochloride as well as 37˚C plates were imaged after two days. Plates containing 50 ng/mL cycloheximide (with or without 5 mM guanidinium hydrochloride), 1.5 mM paromomycin, 50 μg/ml hygromycin B, grown at 30˚C were imaged after three days. Plates with 5 mM guanidinium hydrochloride grown at 37˚C were imaged after six days.

CAT tail aggregation is associated with compromised CAT tail degradation
We began our investigation into the effects of aggregation on CAT tail-mediated degradation by validating the model RQC substrate RQCsub. RQCsub consists of green fluorescent protein (GFP) attached via an inert linker including a tobacco etch virus (TEV) protease site to twelve stall-inducing arginine CGN codons ( Fig 1A) [42]. Translation of RQCsub produces a stalled GFP-linker-arginine "arrest product" [19]. The RQCsub arrest product accumulated at low levels in wt strains by SDS-PAGE ( Fig 1A). The arrest product increased in abundance after we compromised RQC function by deleting either of two RQC genes: the CAT tail-elongating factor RQC2 or the E3 ubiquitin ligase LTN1 (Fig 1A). This result indicates that Rqc2 and Ltn1 contribute to RQCsub degradation. While deletion of RQC2 or LTN1 stabilized RQCsub, the . Right, IB of RQCsub immunoprecipitated (IPed) from RQC2 overexpression and LTN1 deletion (ROLD) lysate. Arrow indicates molecular weight of RQCsub without CAT tails. Asterisk denotes the full-length RQCsub protein product, produced when ribosomes translate through the stall sequence (region past the stall not pictured in schematic). GFP, green fluorescent resultant protein products displayed different mobilities: RQCsub from rqc2Δ cells migrated as a crisp band and that from ltn1Δ cells migrated as a higher molecular weight smear (Fig 1A). This smear collapsed into a single band after additional deletion of RQC2 in the ltn1Δ strain (Fig 1A), consistent with the smear containing CATylated RQCsub. Taken together, these results confirm that RQCsub is degraded via RQC and CATylated by Rqc2.
To vary the aggregation propensity of CAT tails, we overexpressed RQC2, which has previously been shown to increase CAT tail aggregation [38]. Overexpression of RQC2 with the TDH3 promoter increased Rqc2 levels 28.4-fold relative to the native promoter, as assessed by levels of red fluorescent protein (RFP)-tagged Rqc2 (S1 Fig panel A). In the RQC2 overexpression and LTN1 deletion strain (hereafter named "ROLD"), a proportion of RQCsub remained in the well rather than migrating into an SDS-PAGE gel ( Fig 1A) despite prior boiling in dodecylsulfate detergent. This detergent-resistant species was more prominent in ROLD than in ltn1Δ lysates or lysates from cells with RQC2 overexpression alone ("RQC2-OE") (S1 Fig panel  B). TEV treatment, which cleaves RQCsub's C-terminus to remove its CAT tail, enabled RQCsub purified from ROLD cells to migrate as a single band into the gel (S1 Fig panel C). The properties of this detergent-resistant species are consistent with RQC2 overexpression reducing the solubility of CATylated RQCsub in the ltn1Δ background. Microscopy revealed that RQCsub had a diffuse localization in the majority of ltn1Δ cells (<1% of cells with inclusions) but became more punctate in ROLD cells (56.9% of cells with inclusions) (Fig 1B). To test whether these inclusions depended on CATylation, we disrupted CATylation by deleting RQC2 or overexpressing a CATylation-incompetent rqc2 aaa allele [36] instead of RQC2-WT. Neither of these perturbations (rqc2Δ ltn1Δ nor rqc2 aaa -OE ltn1Δ) induced the punctate RQCsub localization we observed in ROLD cells (Fig 1B), confirming that formation of RQCsub inclusions requires CATylation. These results demonstrate that the ROLD genetic background can be used to enhance CAT tail aggregation.
We next sought a tool to inhibit CAT tail aggregation. Millimolar concentrations of the chaotrope guanidinium hydrochloride (GuHCl) in yeast growth medium limit CAT tail aggregation and block propagation of aggregated yeast prions by inhibiting the ATPase activity of the chaperone Hsp104 [39,[43][44][45][46][47][48][49]. When we treated ROLD cells with GuHCl, we observed that RQCsub still localized to inclusions by microscopy ( S1 Fig panel D). However, monitoring RQCsub mobility in SDS-PAGE assessed how GuHCl affected RQCsub more sensitively than microscopy, revealing that GuHCl treatment enabled a larger proportion of RQCsub extracted from ROLD cells to migrate into the resolving gel rather than remaining in the well (Fig 1C). The presence of GuHCl in growth medium thus reduces the detergent insolubility of CAT tails and can thereby serve as a tool to inhibit CAT tail aggregation.
Single-color model RQC substrates like RQCsub, while useful for visualizing CAT tails by SDS-PAGE and microscopy, have limited utility in degradation measurements due to noisy expression [35]. To assess how aggregation affects degradation of CATylated RQC substrates, we utilized the two-color model RQC substrate RQCsub LONG [35] (Fig 1D and 1E). RQCsub-LONG includes an internal expression control, RFP followed by viral T2A peptides, at the N- Left, schematic of expression-controlled model RQC substrate RQCsub LONG . Center, definitions of "stability," "Δ stability," and "CAT tailmediated degradation." (F) and (G) Δ stability measurements of RQCsub LONG expressed in ltn1Δ and ROLD cells after HUL5 deletion to assess CAT tail degron activity. P-values above bars are derived from t-tests for particular contrast, measuring how significantly different HUL5 deletion-induced stabilization is between either ltn1Δ and ROLD backgrounds in F or between different GuHCl concentrations in G. Error bars represent the standard error of the mean (s.e.m.). https://doi.org/10.1371/journal.pone.0227841.g001 terminus of GFP-linker-arginine (Fig 1E). Ribosomes translating RQCsub LONG produce RFP then skip formation of a peptide bond within the T2A sequence [50,51], severing RFP from the GFP-linker-arginine RQC substrate (Fig 1E). As a result, each round of translation produces RFP and GFP stoichiometrically, but only GFP becomes an RQC substrate (Fig 1E). When we analyzed GFP and RFP separately, we observed that each signal was lower in ROLD than in ltn1Δ cells (S2 Fig panel A). This loss of fluorescent signal was especially evident for the expression control RFP (S2 Fig panel A), indicative of reduced RQCsub LONG translation in the ROLD background. This reduction in translation may explain why we observed lower RQCsub levels in the ROLD background compared to ltn1Δ by SDS-PAGE ( Fig 1A) and highlights the need for a precise way to measure degradation that is not confounded by changes in translation.
We then controlled for these differences in translation to quantify how increased CAT tail aggregation (in the ROLD strain) affects degradation of CATylated RQCsub LONG . The ratio of GFP:RFP at the single-cell level serves as an internal, expression-controlled readout of RQCsub LONG stability (Fig 1E). CAT tail-mediated degradation of RQCsub LONG in the absence of Ltn1 requires the proteasome-associated E3 ubiquitin ligase Hul5 and the proteasome [35]. Therefore, comparing stability before and after HUL5 deletion quantifies CAT tailmediated degradation (Fig 1E). RQCsub LONG stability did not change after treatment of ltn1Δ hul5Δ or ROLD hul5Δ cells with the proteasome inhibitor bortezomib (S2 Fig panel B), supporting the requirement for Hul5 in CAT tail-mediated degradation of RQCsub LONG by the proteasome. HUL5 deletion stabilized RQCsub LONG by 30.2% in ltn1Δ and 13.6% in ROLD backgrounds (Fig 1F), consistent with RQC2 overexpression reducing CAT tail-mediated RQCsub LONG degradation. These results demonstrate that the ROLD genetic background, which potentiates CAT tail aggregation, impairs CAT tail-mediated degradation.
To ensure that it is aggregation (and not another effect of ROLD) that reduces CAT tailmediated degradation, we inhibited aggregation with GuHCl and measured RQCsub LONG stability. We observed that GuHCl treatment had two effects on RQCsub LONG degradation in ROLD strains. First, increasing concentrations of GuHCl increased the magnitude of the effect of HUL5 deletion on the stability of RQCsub LONG (from 13.6% to 20.9% to 28.2% at 0, 5, and 10 mM GuHCl respectively; Fig 1G), suggesting that treating ROLD cells with GuHCl rescues degradation of RQCsub LONG . Second, the magnitude of bortezomib-induced RQCsub LONG stabilization increased with increasing GuHCl when Hul5 was intact (S2 Fig panel C). This result indicates that the degradation that GuHCl rescued involved both Hul5 and the proteasome. Taken together, these data imply that the aggregation-conducive ROLD background impairs CAT tail-mediated degradation, and solubilization of CAT tails rescues this impairment. We thereby posit that CAT tail aggregation antagonizes CAT tail-mediated degradation.

CAT tails exert proteotoxic stress in their aggregated state
To determine how aggregation affects the proteotoxicity of CAT tails, we measured proteotoxic stress after varying CAT tail aggregation propensity (Fig 2A). Proteotoxic stress activates the transcription factor Heat Shock Factor 1 (Hsf1), which then orchestrates transcription of chaperones and other proteostasis-restoring factors [52,53]. Hsf1 activation scales with the magnitude of proteotoxic stress and a previously-published reporter can measure the degree of activation [19]. This reporter consists of GFP under the control of an artificial promoter containing four heat shock elements [54], the consensus binding site for Hsf1 (Fig 2B). LTN1 deletion alone activated the Hsf1 reporter 2.07-fold relative to wt, while additional RQC2 overexpression (ROLD) led to 22.6-fold activation (Fig 2B). The potent activation observed in the ROLD background required LTN1 deletion, as RQC2 overexpression in the wt background weakly affected the reporter (Fig 2B). Thus, the combination of RQC2 overexpression and LTN1 deletion, which potentiates CAT tail aggregation (S1 Fig panel B) [38], generates a synthetic interaction that results in elevated Hsf1 activity. This increase in Hsf1 activity is consistent with an increase in proteotoxic stress.
If CAT tail aggregation causes proteotoxicity, we reasoned that solubilization of CATylated proteins by GuHCl treatment might reduce this toxicity. GuHCl treatment activated the Hsf1 reporter in a dose-dependent manner in wt, RQC2-OE, and ltn1Δ strains (Fig 2C), indicating that GuHCl alone exerts some proteotoxic stress in cells. Despite this general activation, GuHCl decreased Hsf1 reporter activation in ROLD cells (Fig 2C), suggesting that solubilization of CAT tails alleviates the Hsf1 activation associated with the ROLD background. We tested the specificity of this response by measuring the effect of GuHCl on the Hsf1 reporter in other genotypes with high Hsf1 activity [19] (Fig 2D). Hsf1 reporter deactivation was unique to the ROLD condition, as GuHCl increased Hsf1 reporter activation in ump1Δ, hsp104Δ, and hsc82Δ strains and had no significant effect in the ssa2Δ strain (Fig 2D). These results suggest that CAT tails exert proteotoxic stress more potently in their aggregated than in their soluble state.
To ensure that our Hsf1 measurements reflect perturbations to proteostasis, we analyzed how CAT tail aggregation affects cells' sensitivity to stressors. At 30˚C, all of the strains we examined (wt, RQC2 overexpression, ltn1Δ, and ROLD) grew equally well in a spot assay in the absence and presence of 5 mM GuHCl (Fig 2E, plates 1 and 2). This result demonstrates that neither the ROLD background nor GuHCl impairs growth in the absence of stress. We then introduced two mild stressors to perturb proteostasis: the thermal stress of 37˚C incubation or introduction of 50 ng/ml cycloheximide to increase translational stalling. Neither of these stresses induced a growth defect in the RQC2-OE or ltn1Δ strains relative to wt (Fig 2E, plates 3 and 5). By contrast, 37˚C incubation and 50 ng/ml cycloheximide severely inhibited growth in the ROLD strain (Fig 2E, plates 3 and 5). Increased CAT tail aggregation thus co-occurs with sensitivity to thermal and ribosome stalling stresses. Additional GuHCl partially rescued growth of the ROLD strain in the presence of cycloheximide (Fig 2E, plate 5 versus plate 6). The combination of GuHCl and 37˚C led to a synthetic growth defect in all strains (Fig 2E, plate 3 versus plate 4). Despite this growth defect, the ROLD strain grew comparably to wt, RQC2-OE, and ltn1Δ strains when exposed to both GuHCl and 37˚C (Fig 2E, plate 4). Thus, CAT tail aggregation sensitizes cells to mild thermal and ribosome stalling stress. The agreement with our Hsf1 measurements (Fig 2B and 2C) suggests that CAT tail aggregation perturbs proteostasis under the conditions we measured rather than facilitating adaptation to stress.

Disruption of CAT tail aggregation by Pol III perturbation restores CAT tail degradation and alleviates proteotoxicity
Given that GuHCl mildly reduced CAT tail aggregation, we sought an orthogonal perturbation that would more potently inhibit CAT tail aggregation. To find such a perturbation, we screened for mutations that reduced Hsf1 reporter activation in the RQC-compromised ltn1Δ rps0aΔ strain, which exhibits Hsf1 hyperactivation [19]. After crossing a genome-wide collection of hypomorphs into this background as well as a wt background via the synthetic genetic array method [41], we found that all reproducible hits other than RQC2 (whose deletion blocks CATylation) were related to RNA Polymerase III (Pol III). These hits included a non-essential polymerase biogenesis factor, BUD27 [55], and the essential Pol III subunit RPC17. Disruption of each of these by deletion or a "decreased abundance by mRNA perturbation" allele (DAmP) [56] nearly eliminated activation of the Hsf1 reporter in ROLD cells compared to wt (S3 Fig  panel A). To assess the specificity of this effect to RQC-related Hsf1 activation, we measured the effect of BUD27 and RPC17 perturbations in strains with elevated Hsf1 signaling (Fig 3A). Perturbations to BUD27 and RPC17 reduced Hsf1 reporter activity in wt, ump1Δ, hsp104Δ, hsc82Δ, and ssa2Δ strains (Fig 3A, black versus grey bars). However, the magnitude of this reduction in each strain was substantially less than the 20.7-fold and 13.7-fold reduction we observed in the ROLD background after disruption of BUD27 and RPC17, respectively ( Fig  3A). Furthermore, Hsf1 reporter activation increased in the Pol III-perturbed backgrounds upon deletion of UMP1, HSP104, HSC82, and SSA2 relative to the Pol III-perturbed background strain (S3 Fig panel B). Taken together, these data indicate that Pol III impairment reduces ROLD-induced Hsf1 activation without generally preventing Hsf1 activation in other genetic backgrounds.
Because Pol III transcribes target genes like 5S rRNA and tRNAs that play roles in translation [57], we wondered whether Pol III perturbation blocks ROLD-induced Hsf1 activation by perturbing CAT tail synthesis. Pol III perturbation could block CAT tail synthesis by preventing ribosome stalling, a necessary prerequisite for the RQC pathway to initiate. We evaluated this possibility using a quantitative stalling reporter similar to one previously developed Aggregation of CAT tails blocks their degradation and causes proteotoxicity [11,12] (S3 Fig panel C). BUD27 deletion did not alleviate stalling induced by (CGN)12 (the stalling sequence in RQCsub) in wt or ROLD strains relative to control (S3 Fig panel C). Thus, the effects of Pol III perturbation on translation should not confound our analysis of RQCsub by preventing ribosomes translating RQCsub from stalling. When we monitored RQCsub mobility by SDS-PAGE to directly observe any changes in CAT tails, we observed that RQCsub expressed in the ltn1Δ bud27Δ and ltn1Δ rpc17-DAmP strains migrated as a smear that was higher in molecular weight compared to the crisp band observed in the CATylation-deficient ltn1Δ rqc2Δ strain and lower in molecular weight than the smear observed in the CATylationcompetent ltn1Δ strain (Fig 3B). In the ROLD background, the portion of RQCsub in the resolving gel migrated similarly whether or not we perturbed BUD27 or RPC17 (Fig 3B). Additionally, perturbation of BUD27 and RPC17 increased the abundance of RQCsub that migrated into the resolving gel (discussed further below) (Fig 3B). Thus, Pol III disruption may decrease the efficiency of CATylation in the ltn1Δ background but not in the ROLD background.
Given its ability to mitigate Hsf1 hyperactivation in the ROLD strain (Fig 3A and S3 Fig  panel A) and increase the amount of RQCsub that migrated into a resolving SDS-PAGE gel (Fig 3B), we investigated whether Pol III perturbation could inhibit CAT tail aggregation similarly to GuHCl. ROLD cells carrying hypomorphic BUD27 or RPC17 alleles formed RQCsub inclusions less frequently (3.57% and 4.37% relative to 56.9% in ROLD), consistent with a reduction in CAT tail aggregation (Figs 1B and 3C). Additionally, BUD27 deletion in the ROLD background increased the mobility of RQCsub by SDS-PAGE, reducing the proportion that remained in the well and increasing the proportion that migrated into the resolving gel (S3 Fig panel D). Pol III perturbation thus serves as another tool, in addition to GuHCl, to inhibit CAT tail aggregation.
Given that GuHCl solubilized CAT tails and rescued their degradation in ROLD cells ( Fig  1C and 1G), we asked whether Pol III perturbation yields the same effect. To quantify CAT tail degradation, we measured RQCsub LONG stability in ROLD and ROLD bud27Δ cells (Fig 3D). The difference in RQCsub LONG stability (GFP:RFP) in each of these backgrounds with Hul5 (degradation intact) and without Hul5 (degradation blocked) defined the degree of degradation [35]. We validated this approach by demonstrating that the proteasome inhibitor bortezomib did not significantly stabilize RQCsub LONG in ROLD hul5Δ and ROLD bud27Δ hul5Δ strains (S3 Fig panel E), indicating that HUL5 deletion blocked degradation in the ROLD and ROLD bud27Δ backgrounds. HUL5 deletion stabilized RQCsub LONG in ROLD bud27Δ more than in ROLD cells (33.7% compared to 13.6%; Fig 3E). In agreement with the effect of GuHCl (Fig 1G), these data demonstrate that limiting CAT tail aggregation by perturbing Pol III restores degradation of CATylated proteins.
We then used Pol III perturbation to analyze how reducing CAT tail aggregation affects proteostasis by determining its effects on growth in the presence of stress ( Fig 3D). As previously published, BUD27 deletion and rpc17-DAmP slightly lowered growth rate at 30˚C in the wt background (Fig 3F and S3 Fig panel F) [58,59]. BUD27 deletion strongly impaired growth at 37˚C in the wt (previously observed in [58]) and ROLD backgrounds (Fig 3F). rpc17-DAmP had no effect on the wt strain but restored growth of the ROLD strain at 37˚C (S3 Fig panel F). Thus, the sensitivity to mild thermal stress caused by BUD27 deletion renders this perturbation unable to rescue heat sensitivity in the ROLD background, but perturbation of Pol III with rpc17-DAmP can suppress the ROLD strain's sensitivity to heat.
Next, we analyzed the effect of Pol III perturbation on stress induced by the presence of translation inhibitors. Despite causing a growth defect in wt cells, BUD27 deletion enabled faster growth of ROLD cells in the presence of cycloheximide (Fig 3F). Unexpectedly, the growth rate of ROLD bud27Δ cells upon cycloheximide exposure was greater than that of the bud27Δ strain (Fig 3F). rpc17-DAmP had no effect on growth of wt cells in cycloheximide, but fully rescued growth of ROLD cells in cycloheximide (S3 Fig panel F). BUD27 deletion also rescued growth of the ROLD strain in the presence of the translation inhibitor paromomycin, but not hygromycin (S3 Fig panel G). However, BUD27 deletion alone strongly sensitized growth of wt yeast to hygromycin (S3 Fig panel G) [58], perhaps explaining why this perturbation failed to rescue growth of hygromycin-exposed ROLD yeast. Pol III perturbation, similarly to GuHCl, can thereby reduces cells' sensitivity to increased ribosome stalling in addition to inhibiting CAT tail aggregation. Taken together, our observations suggest that improving CAT tail solubility via Pol III perturbation can enhance adaptation to thermal and ribosome stall-inducing stress.

Discussion
We have examined how CAT tail aggregation affects the cell's ability to cope when its burden of incomplete proteins exceeds Ltn1's ability to ubiquitylate them. In cells with compromised Ltn1 function, we modulated the aggregation propensity of CAT tails to determine how aggregation affects the cellular capacity to handle elevated levels of incomplete proteins. Aggregation compromised degradation of CATylated proteins ( Fig 1F) and diminished cellular fitness during stress (Fig 2E), while inhibiting aggregation reversed each of these effects (Figs 1G, 2E, 3E and 3F). Our findings suggest that CAT tail aggregation is detrimental, rather than adaptive, in cells with diminished Ltn1 function.
Since their discovery, multiple functions have been attributed to CAT tails. These include a role in assisting Ltn1 activity [27,35,37] and Ltn1-independent roles such as targeting Ltn1-evading RQC substrates for degradation [35] or serving as aggregation-inducing post-translational modifications [38]. We previously presented evidence supporting an adaptive, Ltn1-independent function for CAT tails by showing that loss of CAT tails sensitized cells to increased ribosome stalling (cycloheximide) specifically in the absence of Ltn1 [35]. Here we show that this adaptive Ltn1-independent role is unlikely to be CAT tail-mediated aggregation, as aggregation instead increased sensitivity to cycloheximide (Figs 2E and 3F). Combined with the recent finding that CAT tail degron function is shared with prokaryotic Ala tails [24], our results suggest that marking escaped RQC substrates for degradation is the primary function of CAT tails.
Our findings elevate the question of why yeast and metazoan CAT tails have retained their threonine content, and thereby aggregation propensity, through evolution. We propose two models to explain this. In the first model, CAT tail aggregation facilitates adaptation to an asof-yet untested stress with evolutionary relevance. The tools we used here to control aggregation may prove useful in probing this model by systematically assessing the fitness benefit of CAT tail aggregation in diverse conditions. The second model proposes that the inclusion of threonine in CAT tails facilitates a function of CAT tails other than aggregation. For instance, threonine incorporation could produce more potent CAT tail degrons, assist in Ltn1-mediated ubiquitylation, or aid in mounting a protective heat shock response. To preserve threonine content, the fitness benefit from this threonine-mediated CAT tail function would need to outweigh any fitness deficits arising from CAT tail aggregation. Given the recent discovery of Rqc2 homologs that incorporate different amino acids into CAT tails [24,34], domain-swapping between these homologs to bias CAT tail amino acid content may assist in determining the function of threonine in CAT tails in future studies. Broadly, understanding how the variable amino acid composition of CAT tails has evolved to meet physiological demands in different organisms may prove a fruitful area of research.
We exploited RQC2 overexpression and Pol III disruption as tools to genetically modify CAT tail aggregation. While these proved useful in analyzing CAT tails in their aggregated and un-aggregated states, the underlying mechanism behind these perturbations is not the focus of this study (nor does it influence our conclusions) and remains unclear. RQC2 overexpression has been proposed to enhance aggregation because endogenous Rqc2 is limiting for CATylation [38]. Under such a scenario, RQC2 overexpression would enable CATylation of more escaped RQC substrates that could then nucleate CAT tail aggregates. Although pleiotropic [57], Pol III perturbation could function in the opposite manner. By limiting translation initiation [58], Pol III perturbation could decrease the total amount of CATylated proteins in the cell and thereby disfavor nucleation of CAT tail aggregation. Alternatively, Pol III perturbation may alter the cellular pool of tRNAs and, thereby, the tRNAs that Rqc2 recruits to synthesize CAT tails. Indeed, disruption of Pol III altered the mobility of soluble CAT tails by SDS-PAGE (Fig 3B, lanes 3 and 4 compared to lane 2), indicative of a change in composition. The altered composition of CAT tails under Pol III-perturbed conditions may improve their solubility. Further studies will be needed to decipher the mechanisms by which these perturbations alleviate CAT tail aggregation.
We report here that CAT tail aggregation modifies the toxicity associated with disrupted Ltn1 function. In mice, a hypomorphic LTN1 allele induces a progressive neurodegeneration that shares phenotypic similarities with amyotrophic lateral sclerosis (ALS) [32]. It is possible that degenerating neurons in these LTN1-hypomorphic mice harbor toxic CAT tail aggregates, as has been observed in a Drosophila model of Parkinson Disease [34]. As the connection between the phenotypes in these disease models and human disease becomes more clear, future studies can focus on strategies to mitigate toxicity associated with compromised Ltn1. Our study serves as a proof-of-principle that disrupting CAT tail aggregation lessens this toxicity. Error bars indicate s.e.m. from three independent cultures. P-values from paired t-tests indicated above bars. (B) Flow cytometry of Pol III-perturbed cells containing an integrated reporter for Hsf1 activation. These data are also contained in Fig 3A, but are reordered here to simplify comparisons within two Pol III-perturbed genetic backgrounds. Error bars as in A. (C) Above, schematic of stalling reporter with the same (CGN)12 stalling sequence contained in RQCsub or a non-stalling (Ser-Thr)6 sequence, similar to a reporter used in refs 11 and 12. Below, flow cytometry of stalling and non-stalling reporters expressed in indicated strains. Error bars as in A. (D) IB of lysates containing RQCsub derived from ROLD compared to ROLD bud27Δ. (E) Additional data to support Fig 3E. Stability measurements of RQCsub LONG expressed in indicated strains with bortezomib (btz) treatment to inhibit the proteasome and HUL5 deletion to block CAT tail degron activity. Error bars indicate s.e.m. from three independent cultures. P-values are given above bars. Results of paired t-tests measuring the significance of bortezomib-induced stabilization are indicated with thick lines. The result of a t-test for particular contrast is indicated with thin lines; this assesses how significantly different HUL5 deletion-induced stabilization is in ROLD compared to ROLD bud27Δ. (F and G) Spot assay of yeast strains grown under indicated conditions. (TIF) S1 Table. Yeast strains used in this study. (PDF) S2 Table. Plasmids used in this study. (PDF) S1 Dataset. Results from Hsf1 screen. Each row includes the name of a gene (column 1) that was disrupted with a hypomorphic allele (deletion for non-essential genes and DAmP for essential genes) and measured in a strain with a GFP-fluorescent Hsf1 reporter in either a wt (column 2) or ltn1Δ rps0aΔ background (column 3). The interaction (difference) between these measurements is reported in column 4. (XLS) S1 Raw Images. Uncropped images of immunoblots contained in the main and supplementary figures. (TIF)