A connection between the ribosome and two S. pombe tRNA modification mutants subject to rapid tRNA decay

tRNA modifications are crucial in all organisms to ensure tRNA folding and stability, and accurate translation. In both the yeast Saccharomyces cerevisiae and the evolutionarily distant yeast Schizosaccharomyces pombe, mutants lacking certain tRNA body modifications (outside the anticodon loop) are temperature sensitive due to rapid tRNA decay (RTD) of a subset of hypomodified tRNAs. Here we show that for each of two S. pombe mutants subject to RTD, mutations in ribosomal protein genes suppress the temperature sensitivity without altering tRNA levels. Prior work showed that S. pombe trm8Δ mutants, lacking 7-methylguanosine, were temperature sensitive due to RTD, and that one class of suppressors had mutations in the general amino acid control (GAAC) pathway, which was activated concomitant with RTD, resulting in further tRNA loss. We now find that another class of S. pombe trm8Δ suppressors have mutations in rpl genes, encoding 60S subunit proteins, and that suppression occurs with minimal restoration of tRNA levels and reduced GAAC activation. Furthermore, trm8Δ suppression extends to other mutations in the large or small ribosomal subunit. We also find that S. pombe tan1Δ mutants, lacking 4-acetylcytidine, are temperature sensitive due to RTD, that one class of suppressors have rpl mutations, associated with minimal restoration of tRNA levels, and that suppression extends to other rpl and rps mutations. However, although S. pombe tan1Δ temperature sensitivity is associated with some GAAC activation, suppression by an rpl mutation only modestly inhibits GAAC activation. We propose a model in which ribosomal protein mutations result in reduced ribosome concentrations, leading to both reduced ribosome collisions and a reduced requirement for tRNA, with these effects having different relative importance in trm8Δ and tan1Δ mutants. This model is consistent with our results in S. cerevisiae trm8Δ trm4Δ mutants, known to undergo RTD, fueling speculation that this model applies across eukaryotes.


Introduction
tRNAs contain numerous post-transcriptional modifications.These tRNA modifications are highly conserved evolutionarily within each domain of life, and in some cases throughout all of the domains [1,2].Modifications impart important functional or structural properties to the tRNAs, and consequently, lack of any of a number of tRNA modifications results in growth defects in the budding yeast Saccharomyces cerevisiae and in neurological, mitochondrial, or other disorders in humans [3,4].Modifications within the anticodon loop frequently affect the efficiency or fidelity of translation [5][6][7][8][9][10], consistent with their interactions in the A-site of the ribosome.By contrast, modifications in the main tRNA body often impair the folding and/or stability of the tRNA [11][12][13][14].
The biology of body modifications in eukaryotes is best understood in S. cerevisiae, in which lack of any of several body modifications targets a subset of the hypomodified tRNAs to either or both of two quality control decay pathways, particularly at higher temperatures (see [3]).Lack of each of several modifications, alone or in combination with other body modifications, targets mature tRNAs for 5'-3' exonucleolytic degradation by the rapid tRNA decay (RTD) pathway, which is catalyzed by Xrn1 and Rat1 and inhibited by lack of Met22 [14,15], due to the accumulation of the metabolite adenosine 3',5' bis-phosphate [16,17].Thus, trm8Δ trm4Δ mutants lack 7-methylguanosine at G 46 (m 7 G 46 ) and 5-methylcytidine (m 5 C) and are temperature sensitive due to RTD of tRNA Val(AAC)  [13,14].Similarly, both tan1Δ trm44Δ mutants (lacking 4-acetylcytidine at C 12 (ac 4 C 12 ) and 2'-O-methyluridine at U 44 ) and trm1Δ trm4Δ mutants (lacking N 2 ,N 2 -dimethylguanosine at G 26 (m 2,2 G 26 ) and m 5 C) are temperature sensitive due to RTD of tRNA Ser(CGA) and tRNA Ser(UGA) , and the temperature sensitivity of each of the trm8Δ, trm1Δ, and tan1Δ single mutants is associated with tRNA decay and inhibited by a met22Δ mutation [14,15,18].By contrast, trm6 mutants, lacking m 1 A 58 , are targeted by both the nuclear surveillance pathway and the RTD pathway.The nuclear surveillance pathway degrades pre-tRNA i Met lacking m 1 A 58 using the 3'-5' exonucleases Rrp6 and Rrp44 of the nuclear exosome, after oligoadenylation by Trf4 of the TRAMP complex [12,19,20], and in

PLOS GENETICS
addition, the RTD pathway has a prominent role in decay of tRNA i Met lacking m 1 A 58 in trm6 mutants [21].
Our genetic analysis of the biology of body modification mutants in S. pombe has revealed strong evidence for conservation of the use of the RTD pathway in monitoring tRNA quality across at least the 600 million years separating S. pombe and S. cerevisiae, which pre-dates the Cambrian explosion during which the major animal phyla emerged [22].Thus, we showed that S. pombe trm8Δ mutants, lacking m 7 G 46 , were temperature sensitive due to reduced levels of tRNA Tyr(GUA) , and to some extent tRNA Pro(AGG) , and that each of four trm8Δ suppressors with restored tRNA levels had mutations in the RAT1 ortholog dhp1 [23].Similarly, we found that the temperature sensitivity of S. pombe trm6Δ mutants, lacking m 1 A 58 , was due to the decay of tRNA i Met by the RTD pathway, and not the nuclear surveillance pathway, based on isolation of multiple trm6Δ suppressors with mutations in dhp1 or tol1 (the MET22 ortholog), and the lack of trm6Δ suppression by mutation of the TRF4 ortholog cid14 [21].
Our additional analysis of S. pombe trm8Δ suppressors also led to the discovery of a connection between the onset of RTD and activation of the general amino acid control (GAAC) pathway, equivalent to the integrated stress response pathway in humans.Thus, each of seven S. pombe trm8Δ suppressors that were sensitive to the GAAC activator 3-aminotriazole (3-AT) had null mutations in one of three components of the GAAC pathway (gcn1 + , gcn2 + , or gcn3 + (tif221 + )) and had partially restored levels of tRNA Tyr and tRNA Pro(AGG) , consistent with the interpretation that decay of tRNA Tyr and tRNA Pro(AGG) in S. pombe trm8Δ mutants triggers GAAC activation, causing a further loss of these tRNAs.In support of this interpretation, we found that the lowest temperature at which tRNA decay was detected in S. pombe trm8Δ mutants coincides with the temperature at which a growth defect was first observed and GAAC activation was first evident [23].Moreover, the connection between the onset of RTD and GAAC activation extends to S. cerevisiae, as the lowest temperature at which the well-studied S. cerevisiae trm8Δ trm4Δ mutant showed a growth defect was accompanied by significant tRNA Val(AAC) decay and GAAC activation.However, GAAC activation has opposite consequences in S. cerevisiae trm8Δ trm4Δ mutants, compared to S. pombe trm8Δ mutants, as introduction of GAAC mutations led to exacerbated temperature sensitivity and a further reduction in tRNA Val(AAC) levels, consistent with the GAAC pathway acting to prevent further tRNA loss in S. cerevisiae [23].
Here, we show the existence of a connection between tRNA modification biology in S. pombe and proteins in the ribosome, which also appears to extend to S. cerevisiae.We find that a prominent class of spontaneous suppressors of the temperature sensitivity of S. pombe trm8Δ mutants have mutations in rpl genes, encoding proteins in the large ribosomal subunit, and that suppression is accompanied by little or no restoration of tRNA levels and reduced activation of the GAAC pathway.This suppression of S. pombe trm8Δ mutants extends to each of several tested rpl and rps genes (encoding proteins in the small subunit).Furthermore, the connection between tRNA modification and proteins in the ribosome also extends to another S. pombe body modification mutant.Thus, we find that S. pombe tan1Δ mutants, lacking ac 4 C 12 in their tRNAs, are temperature sensitive due to RTD of two tRNAs, and that a large class of tan1Δ suppressors have mutations in ribosomal protein genes.However, S. pombe tan1Δ mutants have reduced GAAC activation, compared to S. pombe trm8Δ mutants, and an rpl502Δ mutation does not dramatically inhibit GAAC activation, but suppresses the S. pombe tan1Δ growth defect more efficiently than a gcn2Δ mutation.In contrast, we find that rpl and rps mutations exacerbate the temperature sensitivity of the well-studied S. cerevisiae trm8Δ trm4Δ mutant, known to be subject to RTD.We interpret these results in terms of a model in which rpl and rps mutations in both organisms lead to reduced ribosome concentrations [24,25], resulting in a reduced need for tRNA, as well as reduced ribosome collisions, which in turn leads to reduced GAAC activation) [26,27].In this model, the relative importance of these effects on specific modification mutants depends on the particular modification and the organism.Based on these results, we speculate that mutations in ribosomal protein genes will similarly act to influence growth defects of other modification mutants in S. pombe and S. cerevisiae, as well as in other eukaryotes.

A class of trm8Δ suppressors have mutations in rpl genes and distinct growth properties
To further understand mechanisms mediating the temperature sensitivity of S. pombe trm8Δ mutants, we identified additional suppressors with growth phenotypes different from those with mutations in the RTD pathway or the GAAC pathway, and then sequenced their genomes.We tested sensitivity to 5-fluorouracil (5-FU) for two reasons.First, our previous work showed that S. pombe trm8Δ mutants were modestly sensitive to 5-FU at higher temperatures [23], consistent with the 5-FU sensitivity of S. cerevisiae trm8Δ mutants [28].Second, we had shown that S. pombe trm8Δ dhp1 suppressors were 5-FU resistant [23], likely because the mutations in the dhp1 gene (orthologous to S. cerevisiae rat1) protect tRNAs from decay due to the lack of both m 7 G, and to modifications inhibited by 5-FU [29][30][31].We also tested 3-AT sensitivity because S. pombe trm8Δ gcn2 mutants, like other strains with mutations in the GAAC pathway, are sensitive to 3-AT, a competitive inhibitor of His3 that leads to histidine starvation and activation of the GAAC pathway, resulting in massive re-programming of gene expression that includes increased biosynthesis of His3 [32].
In this way, we identified a group of four trm8Δ suppressors that suppressed trm8Δ temperature sensitivity in rich (YES) and complete minimal media lacking histidine (EMMC-His) almost as efficiently as a trm8Δ suppressor with a representative dhp1 mutation (trm8Δ dhp1-W326L) or GAAC mutation (trm8Δ gcn2-M1I) [23].However, unlike the dhp1 suppressors, these suppressors grew poorly on media containing 5-FU at 33˚C, and unlike the GAAC or dhp1 suppressors, these suppressors were resistant to 3-AT at 37˚C (Figs 1A and S1).This class of trm8Δ suppressors all proved to have mutations in rpl genes (rpl1701-Q72X, rpl502-Y44X, rpl1502-X201E, and rpl1102-K93X), encoding different proteins (Rpl17, Rpl5, Rpl15, and Rpl11) in the 60S subunit of the ribosome.Based on Pombase, all of these mutations are in one of two paralogous genes and are not essential, and the three mutant alleles with a premature stop codon are predicted to be null mutants.
To determine if the rpl mutations were responsible for the restoration of growth in the trm8Δ rpl suppressor mutants, we randomly selected two of them to test for complementation by a [leu2 + ] plasmid containing the corresponding rpl + gene.Consistent with complementation, we found that, like the trm8Δ [leu2 + ] strain, expression of rpl502 + in the trm8Δ rpl502 mutant, or of rpl1701 + in the trm8Δ rpl1701 mutant (but not the vector controls) resulted in temperature sensitivity in EMMC-Leu media at 39˚C, 5-FU resistance at 33˚C, and 3-AT sensitivity at 38˚C (S2 Fig) .Based on these results, and on the similar growth properties of all four trm8Δ rpl suppressor strains, it seems highly likely that the corresponding rpl mutations are responsible for suppression in the other two trm8Δ rpl suppressor strains.

Suppression of the S. pombe trm8Δ temperature sensitivity by each of three rpl mutations occurs without significant restoration of tRNA levels
As the temperature sensitivity of S. pombe trm8Δ mutants is due to decay of tRNA Tyr(GUA) , and to a limited extent tRNA Pro(AGG)  [23], we analyzed tRNA levels of trm8Δ rpl suppressors after temperature shift from 30˚C to 38.5˚C in EMMC-Leu media.As we did previously [23], we normalized measured tRNA levels to tRNA levels of the non-Trm8 substrate tRNA Gly(GCC) , and then to levels in WT strains at 30˚C.As expected, in trm8Δ mutants at 38.5˚C, relative tRNA Tyr(GUA) levels were reduced, to 30% of WT levels (Fig 1B and 1C).Surprisingly, we found that in the trm8Δ rpl1701-Q72X suppressor the tRNA Tyr(GUA) levels were not restored, but were if anything slightly reduced, relative to the trm8Δ mutant (23% vs 30%).Similarly, the tRNA Pro(AGG) levels of the trm8Δ rpl1701-Q72X mutant (8%) were similar to, or slightly less, than those of the trm8Δ mutant (9%), whereas levels of a known unaffected Trm8 substrate (tRNA Thr(AGU) ) remained constant.Thus, for the trm8Δ rpl1701-Q72X suppressor, suppression is occurring without a detectable increase in levels of tRNA Tyr(GUA) or of tRNA Pro(AGG) .Similar analysis shows that the trm8Δ rpl502-Y44X suppressor had slightly increased levels of tRNA Tyr(GUA) (36% vs 30%) and of tRNA Pro(AGG) (13% vs 9%) relative to WT (Fig 1B and 1C), and a follow-up analysis shows that the trm8Δ rpl1102-K93X suppressor also did not restore the reduced levels of tRNA Tyr(GUA) and tRNA Pro(AGG) , whereas levels of a known unaffected Trm8 substrate (tRNA Val(AAC) ) remained constant (S3 Fig) .By contrast, our prior results showed that each of four independent trm8Δ dhp1 suppressors invariably restored tRNA Tyr(GUA) levels by more than 2.8-fold (2.8 to 3.9 fold) relative to those in the trm8Δ mutant, to more than 72% (73-85%) of WT levels, whereas GAAC mutations partially restored tRNA Tyr(GUA) levels (by 1.7 to 2.1 fold), to 35-58% of WT levels [23].Thus, we conclude that as a group, these three rpl mutations suppress the temperature sensitivity of S. pombe trm8Δ mutants without obvious restoration of tRNA Tyr(GUA) levels, or of tRNA Pro(AGG) levels.

Mutations in rpl genes prevent GAAC activation of S. pombe trm8Δ mutants
As we had previously shown that the temperature sensitivity of S. pombe trm8Δ mutants coincided with activation of the GAAC pathway and further loss of tRNA Tyr(GUA)  [23], we measured GAAC activation in two of the S. pombe trm8Δ rpl mutants (trm8Δ rpl1701-Q72X and trm8Δ rpl502-Y44X) at 38.5˚C, using the same bulk RNA that we used for tRNA analysis in Fig 1B and  1C.Consistent with our previous analysis [23], S. pombe trm8Δ mutants at 38.5˚C had a 5.8-fold increase in relative mRNA levels of the GAAC target lys4 + (normalized to the control act1 + mRNA), compared to that from WT cells at 38.5˚C (5.13 vs 0.88) and a 7.5-fold increase compared to that from trm8Δ mutants at 30˚C (5.13 vs 0.68).By contrast, trm8Δ rpl1701-Q72X and trm8Δ rpl502-Y44X mutants had near baseline relative lys4 + expression at both temperatures (0.97 and 0.88 vs 0.88 for WT, at 38.5˚C) (Fig 1D).Similar results were obtained by measuring relative aro8 + mRNA levels, although the GAAC activation is not as strong (S4 Fig), as we observed previously [23].We conclude that the rpl mutations inhibit the GAAC activation normally observed in trm8Δ mutants at 38.5˚C, likely due to reduced ribosome concentrations arising from mutation of one of the two paralogs of each ribosomal protein.
rpl1701-Q72X and trm8Δ rpl502-Y44X suppressors do not appreciably prevent the decay of tRNA Tyr(GUA) and tRNA Pro(AGG) at 38.5˚C in EMMC-Leu media.Strains transformed with a leu2 + vector were grown in triplicate in EMMC-Leu media at 30˚C, diluted into fresh media at 30˚C and 38.5˚C and grown for 10 hours, and then bulk RNA was analyzed for tRNA levels by northern blot analysis as described in Materials and Methods, with the indicated probes.(C).Quantification of tRNA levels of WT, trm8Δ, and trm8Δ rpl mutants at 38.5˚C in Fig 1B .The bar chart depicts relative levels of tRNA species at 30˚C and 38.5˚C, normalized to their levels in the WT strain at 30˚C (each value itself first normalized to levels of the control non-Trm8 substrate tRNA Gly(GCC) ).Standard deviations for each tRNA measurement are indicated.dark shades, 30˚C growth; light shades, 38.5˚C (D).S. pombe trm8Δ rpl1701-Q72X and trm8Δ rpl502-Y44X mutants have reduced GAAC activation at 38.5˚C, relative to trm8Δ strains.Bulk RNA from the growth in Fig 1B was used for RT-qPCR analysis of the mRNA levels of the GAAC target lys4 + and the control act1 + as described in Materials and Methods, and mRNA levels of lys4 + were normalized to those of act1 + .https://doi.org/10.1371/journal.pgen.1011146.g001

A reconstructed S. pombe trm8Δ rpl502Δ mutant has the same properties as a trm8Δ rpl502-Y44X mutant
To ensure that the effects of the trm8Δ rpl502-Y44X mutant were solely due to the rpl502 mutation, rather than some background mutation that arose in the original trm8Δ strain before selection for suppressors, we introduced an rpl502Δ mutation into a clean trm8Δ mutant and tested the reconstructed trm8Δ rpl502Δ mutant.We found that a reconstructed trm8Δ rpl502Δ strain suppressed the growth defect of a trm8Δ mutant almost identically to the trm8Δ rpl502-Y44X mutant, with similar growth in YES and EMMC media at high temperature, and similar 5-FU sensitivity and 3-AT resistance (Fig 2A).Also like trm8Δ rpl502-Y44X mutants, trm8Δ rpl502Δ mutants did not restore tRNA Tyr(GUA) levels after temperature shift to 38.5˚C in YES media (33% vs 41% for trm8Δ mutants) whereas levels of tRNA Tyr(GUA) were modestly restored in trm8Δ gcn2Δ mutants (58% vs 41%), as we observed previously [23] (Fig 2B and  2C).Similarly, tRNA Pro(AGG) levels were not efficiently restored in trm8Δ rpl502Δ mutants (38%) relative to 25% in trm8Δ mutants and 49% in trm8Δ gcn2Δ strains.Furthermore, like trm8Δ rpl502-Y44X mutants, trm8Δ rpl502Δ mutants efficiently suppressed the GAAC activation observed in trm8Δ mutants, measured by relative lys4 + levels (from a 14.5-fold increase in trm8Δ mutants, to a 1.9-fold increase in trm8Δ rpl502Δ mutants), compared to a near baseline 1.2-fold increase in a trm8Δ gcn2Δ strain (Figs 2D and S5).We conclude that the suppression effects of the trm8Δ rpl502-Y44X mutant are explicitly due to the rpl mutation, and infer that the suppression by the trm8Δ rpl1701-Q72X and trm8Δ rpl1102-K93X mutations is also due to the rpl mutations.
We note that, as the control rpl502Δ mutants are as resistant to 3-AT as the trm8Δ rpl502Δ mutants (Fig 2A ), this property is due to the rplΔ mutation and not the trm8Δ mutation in these strains.

Deletion of any of several rpl + or rps + genes suppress the temperature sensitivity of S. pombe trm8Δ mutants
To determine if the spontaneous S. pombe trm8Δ rpl suppressors that we had found by selection represented a specific subset of ribosomal protein genes, we investigated whether or not suppression would extend to deletion of other rpl + or rps + genes.We focused on six ribosomal protein genes that had two copies and thus were not essential, including rpl1701, for which we already had a spontaneous trm8Δ suppressor.Deletion of each of the two rps genes (rps2801Δ and rps802Δ, encoding Rps28 and Rps8) and four rpl genes examined (rpl1601Δ, rpl1202Δ, rpl2802Δ, and rpl1701Δ, encoding Rpl16, Rpl12, Rpl28, and Rpl17) resulted in efficient suppression of the trm8Δ temperature sensitivity in EMMC-His media, although suppression was somewhat weaker for the rpl1601Δ mutation (Fig 3A and 3B).Similarly, all five of the trm8Δ rplΔ and trm8Δ rpsΔ mutants that were robust trm8Δ suppressors were also sensitive on YES +5-FU media at 33˚C, whereas the weak trm8Δ rpl1601Δ suppressor was slightly sensitive on YES+5-FU.We conclude that rplΔ and rpsΔ mutations generally cause suppression of the S. pombe trm8Δ growth defect, presumably as a consequence of the reduced concentrations of the corresponding ribosomal subunits [25,33].
By contrast, we found that 3-AT resistance was not correlated with the efficiency of suppression in trm8Δ rplΔ strain, as one of the robust trm8Δ suppressors was almost completely 3-AT sensitive (trm8Δ rpl1202Δ) and two other robust suppressors were only moderately 3-AT resistant (trm8Δ rps2801Δ and trm8Δ rps802Δ), whereas the robust trm8Δ rplΔ2802Δ and trm8Δ rpl1701Δ suppressor strains were strongly 3-AT resistant.This result is consistent with our finding above that 3-AT resistance is a property of the mutation in the ribosomal protein gene, and not the trm8Δ mutation, However, it is not immediately clear why some mutations in ribosomal protein genes result in 3-AT resistance and others result in sensitivity, although it is known that different mutations in ribosomal protein genes in S. cerevisiae result in different growth properties and phenotypes [24,25].
An S. pombe tan1Δ mutant lacks 4-acetylcytidine and is temperature sensitive due to decay of tRNA Leu(AAG) and tRNA Leu(UAG) by the RTD pathway To explore the generality of the effect of the ribosome and GAAC mutations on the growth of S. pombe modification mutants, we examined an S. pombe tan1Δ mutant.It is well established from work in S. cerevisiae that tan1Δ mutants lack 4-acetylcytidine at C 12 (ac 4 C 12 ) in their tRNA Leu and tRNA Ser species, which have a long variable arm [34], that tan1Δ trm44Δ mutants are temperature sensitive due to RTD of tRNA Ser(CGA) and tRNA Ser(UGA) [14], and that the temperature sensitivity of tan1Δ single mutants is associated with decay of tRNA Ser(CGA) and tRNA Ser(UGA) and is suppressed by a met22Δ mutation [15].Indeed, we find that the temperature sensitivity of S. cerevisiae tan1Δ mutants is completely suppressed by overexpression of tRNA Ser(CGA) and nearly as efficiently suppressed by overexpression of tRNA Ser(UGA) (S6 Fig) .Consistent with the biology of Tan1 in S. cerevisiae, we find that S. pombe tan1Δ mutants are temperature sensitive due to decay of two tRNA species by the RTD pathway.As anticipated based on the conservation of Tan1 sequence and the Tan1 requirement for ac 4 C 12 modification of tRNA Leu and tRNA Ser within a CCG motif [35][36][37][38], purified tRNA Leu(CAA) from S. pombe tan1Δ mutants lacks any detectable ac 4 C, compared to that in WT cells, but has similar levels of four control modifications (Fig 4A).In addition, we find that S. pombe tan1Δ mutants are temperature sensitive on both YES media and EMMC-Leu media at 38˚C (Fig 4B).The S. pombe tan1Δ temperature sensitivity in EMMC-Leu media is associated with reduced levels of tRNA Leu(AAG) and tRNA Leu(UAG) at 30˚C that are further reduced at 39˚C (relative to those in WT cells), whereas levels of each of the other tRNA Leu and tRNA Ser species do not conform to this pattern (Figs 4C, S7, and S8).Furthermore, the temperature sensitivity of S. pombe tan1Δ mutants is almost entirely due to tRNA Leu(AAG) , and to some extent tRNA Leu(UAG) , as overexpression of these two tRNAs, or tRNA Leu(AAG) alone, efficiently suppresses the temperature sensitivity (Fig 4D).Moreover, as anticipated from Tan1 biology in S. cerevisiae, whole genome sequencing resulted in identification of three S. pombe tan1Δ suppressors with dhp1 mutations (dhp1-I196T, dhp1-P677S, and dhp1-P203L), indicating decay of these tRNAs by the RTD pathway; these suppressors will be discussed elsewhere.

A prominent class of S. pombe tan1Δ suppressors have mutations in rpl genes or related genes predicted to reduce 60S subunits
Consistent with our results with S. pombe trm8Δ suppressors, we find that several spontaneous suppressors of the temperature sensitivity of S. pombe tan1Δ mutants have mutations in ribosomal protein genes.Thus, of seven tan1Δ suppressors that are resistant to 3-AT at 35˚C and somewhat sensitive to 5-FU at 38˚C, whole genome sequencing showed that six have rpl mutations (rpl1101-N7fs, rpl1102-R133fs, rpl1502-R67_R68_insVR, rpl1701-T35fs, rpl2802-G113fs, and rpl3001-N80fs) and the seventh has a grn1-N37fs mutations.S. pombe Grn1 and its S. cerevisiae ortholog Nug1 are critical GTPases required for pre-60S rRNA processing [39,40].It is therefore clear that this selection (temperature resistance) and screening procedure (3-ATresistance, 5-FU sensitivity) is consistently yielding tan1Δ suppressors with mutations predicted to reduce the amount of the 60S subunit, just as we observed with trm8Δ suppressors.
We confirmed two of the ribosomal protein mutations as tan1Δ suppressors by introduction of the corresponding deletion mutation into a clean tan1Δ background.Thus, we find and tRNA Leu(UAG) .(A). S. pombe tan1Δ mutants lack ac 4 C in their tRNA Leu(CAA) .S. pombe WT and tan1Δ mutants were grown in triplicate in YES media at 30˚C to late log phase, and tRNA Leu(CAA) was purified from bulk RNA, digested to nucleosides, and analyzed for modifications by HPLC as described in Materials and Methods.(B).S. pombe tan1Δ mutants are temperature sensitive.S. pombe WT and tan1Δ mutants were transformed with a [leu2 + ] that an rpl2802Δ mutation efficiently suppresses the growth phenotypes of an S. pombe tan1Δ mutant, resulting in temperature resistance on EMMC-His media and YES media at 38˚C, 3-AT resistance at 35˚C, and 5-FU sensitivity at 35˚C (Fig 5A).Similarly, an rpl1701Δ mutation also efficiently suppresses each of the growth phenotypes of an S. pombe tan1Δ mutant, resulting in temperature resistance, 3-AT resistance, and 5-FU sensitivity (Fig 5B).
In addition, we find that tan1Δ suppression is not restricted to the specific ribosomal protein mutations that were selected, based on analysis of two other ribosomal protein genes.Thus, we find that an rps23Δ mutation and an rpl502Δ mutation each efficiently suppress the temperature sensitivity of the tan1Δ mutant, although we note that the tan1Δ rps23Δ mutant is not 5-FU sensitive, and that the tan1Δ rpl502Δ mutant is only slightly 3-AT resistant at 35˚C (Fig 5B ).

The rpl502Δ mutation suppresses the temperature sensitivity of the S. pombe tan1Δ mutant without restoring tRNA levels or affecting GAAC activation
To analyze the effects of the tan1Δ rpl mutants on tRNA levels and GAAC activation, we focused on the tan1Δ rpl502Δ mutant, so as to directly compare our results with those of the trm8Δ rpl502Δ mutant.To assess the role of the rpl502Δ mutation in suppression of the S. pombe tan1Δ growth defect, we grew WT, tan1Δ, and tan1Δ rpl502Δ strains at 30˚C in YES media, shifted to 39˚C, and analyzed both tRNA levels and GAAC activation after 9 hours.As we saw for the trm8Δ rpl502Δ strain, we find that the tan1Δ rpl502Δ strain does not have restored tRNA levels after growth at 39˚C (Fig 6A and 6B).Relative tRNA Leu(AAG) levels are reduced substantially in tan1Δ strains after growth at 39˚C (to 24% of WT), and remain low in the tan1Δ rpl502Δ strain at 39˚C (19% of WT), but each of two control tRNAs are unaffected in both strains.However, we find that although an S. pombe tan1Δ mutant activates the GAAC pathway, based on analysis of relative lys4 + mRNA levels, the activation is weak, and is only modestly reduced in a tan1Δ rpl502Δ strain (Fig 6C).Thus, we find that a tan1Δ mutant has a 3.3-fold increase in relative lys4 + mRNA levels at 39˚C (1.62) compared to WT at 39˚C (0.49) and a 2.1-fold increase compared to tan1Δ mutants at 30˚C (0.80).This is substantially weaker than the GAAC activation of a trm8Δ mutant in the same experiment, which increases 23.4-fold compared to WT at 39˚C (11.45 vs 0.49) and 12.6-fold, compared to trm8Δ mutants at 30˚C (11.45 vs 0.91).Moreover, we find that an rpl502Δ mutation has only a minimal effect on GAAC activation by a tan1Δ mutant.Thus, a tan1Δ rpl502Δ mutant has a 2.5-fold increase in relative lys4 + levels at 39˚C relative to 30˚C (1.13 vs 0.46), compared to 2.1-fold for the tan1Δ mutant.and a 2.31-fold increase at 39˚C relative to WT at 39˚C (1.13 vs 0.49), compared to 3.31-fold for the tan1Δ mutant.
A similar analysis after growth of strains in EMMC-Leu media yields similar results regarding the effect of the rpl502Δ mutation on S. pombe tan1Δ mutants (S9 Fig) .We observe reduced levels of tRNA Leu(AAG) in the S. pombe tan1Δ mutant at 39˚C (23% of WT levels at or a [leu2 + tan1 + ] plasmid as indicated, and transformants were grown overnight in EMMC-Leu media, serially diluted, and plated on media as indicated.(C).S. pombe tan1Δ mutants have reduced levels of tRNA Leu(AAG) and tRNA Leu(UAG) at 39˚C. S. pombe WT and tan1Δ mutants were transformed with a [leu2 + ] plasmid, and transformants were grown in EMMC-Leu media to mid-log phase, diluted and shifted to 39˚C and grown for 12 hours, and RNA was isolated at 0, 3, 6, 9, and 12 hours and analyzed by Northern blot as described in Materials and Methods, with the indicated probes.The full set of Northern blot data and quantification is shown in S7 and S8 Figs.shades of blue, WT strains, analyzed at 0 (darkest) through 12 hours (lightest); shades of green, tan1Δ strains (D).The temperature sensitivity of S. pombe tan1Δ mutants is efficiently suppressed by overexpression of tRNA Leu(AAG) and, to some extent, tRNA Leu(UAG) .S. pombe WT and tan1Δ mutants were transformed with a [leu2 + ] plasmid expressing tRNAs as indicated, or a vector control, and transformants were grown overnight in EMMC-Leu media at 30˚C and analyzed for growth, as in Fig 1A, on plates containing EMMC-Leu or YES media.https://doi.org/10.1371/journal.pgen.1011146.g00430˚C), which remain at similar or slightly lower levels in the tan1Δ rpl502Δ strain (19%).Although we observe some activation of the GAAC pathway in an S. pombe tan1Δ mutant, based on analysis of relative lys4 + mRNA levels, the rpl502Δ mutation has only a modest effect on this activation.The tan1Δ mutant has a 3.8-fold increase in relative lys4 + mRNA levels, compared to WT at 39˚C (4.79 vs 1.27) (which is smaller than the 5.7-fold in the trm8Δ mutant), whereas the tan1Δ rpl502Δ strain has a 2.6 fold increase (S9 Fig).These weak effects of the tan1Δ mutant and the tan1Δ rpl502Δ mutant on GAAC activation are consistent with the observation that an rpl502Δ mutation is a more efficient suppressor of the temperature sensitivity of an S. pombe tan1Δ mutant than is a gcn2Δ mutation, on either YES media or EMMC-His media at 38˚C (Fig 6D ).
We infer from these results that the rpl502Δ mutation suppresses that tan1Δ mutant by reducing the need for tRNA Leu(AAG) , and likely tRNA Leu(UAG) , due to the reduced number of translating ribosomes arising from the rpl502Δ mutation, rather than by effects on the GAAC pathway due to reduced ribosome collisions.

S. pombe rpl and rps mutants grow slowly and have altered polysome profiles, indicative of reduced translation
To evaluate the interpretation that rpl and rps mutations suppress the temperature sensitivity of S. pombe trm8Δ and S. pombe tan1Δ mutants due to reduced numbers of translating ribosomes, we explicitly examined mutants for growth and translation properties.It is well documented that growth rates in E. coli and S. cerevisiae are directly correlated with the number of ribosomes and with overall translation in complete growth conditions [25,41].We therefore compared the growth rate of S. pombe WT strains in liquid YES media with that of several otherwise WT strains with rplΔ or rpsΔ mutations, deletion of each of which caused suppression of the temperature sensitivity of S. pombe trm8Δ and/or tan1Δ mutants, as shown above.Consistent with a defect in translation due to loss of ribosomal protein paralogs, we find that each of four rplΔ mutants and three rpsΔ mutants has a substantially reduced growth rate (increased generation time) relative to that of the WT strain in YES media at 30˚C (Fig 7A and 7B and S1 Table).For the rpl502Δ mutant, we also find a substantially increased generation time in liquid YES media at 38.5˚C (S10A Fig) , and a pie streak test, shows that all of the rplΔ and rpsΔ mutants have reduced growth rates on YES plates at 38˚C, although the rpl1202Δ mutant grows nearly as well as WT (S10B Fig).
Polysome profiling of three of these mutants provides additional evidence for a defect in translation in the mutants (Fig 7C).Polysome profiles of S. pombe WT cells in YES media have typical patterns of 40S and 60S subunits, 80S monosomes, and polysomes [42].By contrast, the rpl502Δ and rpl2802Δ mutants each display distinctly lighter polysomes, which tail off rapidly in contrast to the rising polysome peak in WT, indicative of reduced translation [43].Moreover, both rplΔ mutant profiles, but not the WT, exhibit substantial populations of halfmers, monosomes and polysomes with an additional 40S subunit [44], which are a consequence of the reduced numbers of 60S subunits, relative to 40S subunits, and slowed subunit joining during translation initiation [44][45][46][47][48].The rps23Δ mutant also has reduced polysome density, indicative of reduced translation, accompanied by an undetectable 40S peak and a greatly pronounced 60S peak that substantially overshadows the 80 monosome peak [46].Thus, we conclude that rplΔ and rpsΔ mutants have reduced translation in accordance with their growth defects, as they do in S. cerevisiae [25].

In S. cerevisiae, deletion of rpl or rps genes exacerbates the temperaturesensitivity of trm8Δ trm4Δ mutants
To investigate the evolutionary conservation of the effect of loss of ribosomal protein genes on mutants lacking m 7 G, we examined the growth properties arising from deletion of RPL or RPS genes in an S. cerevisiae trm8Δ trm4Δ mutant.This mutant is known to be highly temperature sensitive due to decay of tRNA Val(AAC) by the RTD pathway [13,14], and like the S. pombe trm8Δ mutant, undergoes GAAC activation coincident with temperature sensitivity and the onset of tRNA decay [23].However, unlike the case with S. pombe trm8Δ mutants, gcn2Δ and  (A).An S. pombe rpl502Δ strain grows distinctly slower that a WT strain.An S. pombe rpl502Δ mutant and WT strain were grown in liquid YES media at 30˚C to mid-log phase (OD 600 0.24 and 0.294 respectively), diluted to OD 600 ~0.1 in the same media, and growth was monitored at 30˚C every hour for 8 hours to monitor the growth rate.Data are fit to an exponential, with R 2 values as shown, and generation times determined as ln2/exponent*60 min.(B).Each of several S. pombe rplΔ and rpsΔ strains have growth defects.S. pombe WT, rplΔ, and rpsΔ strains as indicated were grown in triplicate in liquid YES media at 30˚C to mid-log phase (OD 600 range from 0.24 to 0.54), and diluted and grown as described in (A) to derive generation times and R 2 values from exponential fits (S1 Table ).(C). S. pombe rpl502Δ, rpl2802Δ, and rps23Δ strains have polysome profiles indicative of reduced 60S or 40S subunits, and reduced translation.Strains as indicated were grown in YES media at 30˚C to mid-log phase, and cell lysates were analyzed by polysome profiling in a 10-50% sucrose gradient, as described in Materials and Methods.Wild-type lysate is compared with lysates from rpl502Δ, rpl2802Δ, and rsp23Δ mutant strains.Peaks corresponding to halfmer polysomes are indicated by arrows.https://doi.org/10.1371/journal.pgen.1011146.g007gcn1Δ mutations each exacerbate the temperature sensitivity of S. cerevisiae trm8Δ trm4Δ mutants, and exacerbate the loss of tRNA Val(AAC) at 32˚C, leading to the conclusion that GAAC activation in S. cerevisiae trm8Δ trm4Δ mutants prevents further loss of tRNA Val(AAC) and is beneficial [23].
We find that deletion of each of four tested RPL genes (RPL1A, RPL11B, RPL17B, or RPL33B) or three RPS genes (RPS0A, RPS22A, RPS28A) in the S. cerevisiae trm8Δ trm4Δ mutant does not suppress the temperature sensitivity of the strain, but rather, exacerbates its temperature sensitivity on rich (YPD) media and/or minimal (SD) media (Fig 8A and 8B).We note that the exacerbated growth defect in the S. cerevisiae trm8Δ trm4Δ rpl1AΔ mutant is less distinct, which is likely related to the fact that although the tested ribosomal protein genes are all duplicated, the RPL1A gene is known to be the minor paralog [49,50].Overall, the exacerbated growth defect in S. cerevisiae trm8Δ trm4Δ mutants bearing rplΔ or rpsΔ mutations is consistent with the opposite effects of GAAC activation in S. pombe trm8Δ mutants and S. cerevisiae trm8Δ trm4Δ mutants.

Discussion
We have provided evidence here that for each of two S. pombe body modification mutants (trm8Δ mutants and tan1Δ mutants), that are temperature sensitive due to RTD of specific hypomodified tRNAs, mutations in ribosomal protein genes suppress the growth defect without altering tRNA levels.
The finding that S. pombe tan1Δ mutants are temperature sensitive due to RTD is consistent with our work in S. cerevisiae showing that tan1Δ mutants are temperature sensitive due to RTD [14,15,18] (S6 Fig) , and emphasizes the conservation of this aspect of Tan1 biology across 600 million years [22].Recently, several individuals with a specific syndromic neurodevelopmental disorder were found to have bi-allelic mutations in the human TAN1 ortholog THUMPD1, which was linked to complete loss of ac 4 C in small RNA and in purified tRNA Ser(CGA) for one set of derived lymphoblastoid cell lines, and for each of two human KO cell lines [37].Based on our results, one plausible explanation for the neurological disorder of patients with bi-allelic THUMPD1 mutations is that one or more tRNAs is destabilized due to the loss of ac 4 C, resulting in degradation of those tRNAs by the RTD pathway.
We interpret the results presented here as an expansion of our prior model to explain the common link between the RTD pathway and GAAC activation that we observe in both S. pombe trm8Δ mutants and S. cerevisiae trm8Δ trm4Δ mutants (Fig 9).In the original model [23], increased temperature in both of these mutant strains concomitantly reduces the growth rate and triggers decay of specific hypomodified tRNAs (tRNA Tyr(GUA) and tRNA Pro(AGG) in S. pombe trm8Δ mutants, and tRNA Val(AAC) in S. cerevisiae trm8Δ trm4Δ mutants), which in turn activates the GAAC pathway.In S. pombe trm8Δ mutants, GAAC activation results in further loss of the tRNA Tyr(GUA) and tRNA Pro(AGG) , explaining why mutation of GAAC components suppresses their temperature sensitivity.In contrast, in S. cerevisiae trm8Δ trm4Δ mutants, GAAC activation prevents further loss of the tRNA Val(AAC) , explaining why mutation of GAAC components exacerbates their temperature sensitivity and loss of tRNA Val(AAC) .Although it is unclear why mutation of GAAC components has opposite effects on tRNA loss in S. pombe trm8Δ mutants and S. cerevisiae trm8Δ trm4Δ mutants, we have speculated that this was in some way due to the distinctly different battery of transcriptional responses among the 500 or more genes affected during GAAC activation in S. pombe and S. cerevisiae, or to differences in regulation of the activity or levels of RTD components or regulators [23,51].
In the expanded model (Fig 9), the rpl and rps mutations lead to a reduced concentration of ribosomes in both S. pombe and S. cerevisiae, consistent with the fact that many S. cerevisiae rpl or rps mutants result in reduced concentrations of 60S or 40S subunits and reduced translation [25,33], and with our analysis of growth rates and polysome profiles in several S. pombe rplΔ and rpsΔ strains.The reduced population of ribosomes is expected to have two consequences.First, there would be a reduction in collisions between translating ribosomes.As ribosome collisions are known to activate the GAAC (integrated stress response) pathway in S. cerevisiae and mammals [26,27,52], a reduction in collisions is likely to reduce GAAC activation.This reduced GAAC activation is seen in S. pombe trm8Δ rpl mutants, leading to suppression of the temperature sensitivity of S. pombe trm8Δ mutants, and is proposed to occur in S. cerevisiae trm8Δ trm4Δ rplΔ mutants, explaining their exacerbation of the temperature sensitivity of S. cerevisiae trm8Δ trm4Δ mutants.Second, the reduced population of ribosomes in strains with rplΔ and rpsΔ mutations results in reduced translation [25,43], which we propose results in a reduced requirement for tRNA.The reduced requirement for tRNA explains our results in S. pombe tan1Δ mutants, in which an rpl502Δ mutation does not restore tRNA Leu(AAG) levels at 39˚C, and does not significantly inhibit the relatively weak GAAC activation observed in tan1Δ mutants, but does suppress the tan1Δ growth defect more efficiently than a gcn2Δ mutation.

PLOS GENETICS
We infer that the ribosomal protein mutations result in a reduced number of ribosomes, rather than resulting in a specific alteration of ribosome function, for two reasons.First there is large diversity in the rpl and rps mutations/deletions that suppress the growth defect of S. pombe trm8Δ mutants (encoding seven different large subunit proteins and two different small subunit proteins) or of S. pombe tan1Δ mutants (encoding six different large subunit proteins and one small subunit protein, including one from each subunit that was not observed or tested in trm8Δ mutants).Second, these eight different proteins of the large subunit and three different proteins of the small subunit are found at different locations on the mature ribosome [53,54].The isolation of the grn1-N37fs mutation as a tan1Δ suppressor also implicates the concentration of the 60S subunit, given its known crucial role in 60S biogenesis [39,40].
It is curious that many of the spontaneous S. pombe trm8Δ and tan1Δ suppressors that we sequenced after screening for 3-AT resistance and 5-FU sensitivity had mutations in rpl genes or in the grn1 gene (implicated in 60S production), but not in rps genes.Although we do not know the source of this bias, results of others show distinct differences in the growth rates of In each organism, the reduced levels of the corresponding tRNA result in increased ribosome collisions and activation of the GAAC pathway (blue), leading to further loss of tRNA Tyr(GUA) in S. pombe and prevention of further loss of tRNA Val(AAC) in S. cerevisiae [23].In both organisms, the rpl and rps mutations (brown) result in a reduced concentration of ribosomes, leading to inhibition of ribosome collisions and reduced GAAC activation in S. pombe, and presumed to reduce GAAC activation in S. cerevisiae.In both organisms, the reduced concentration of ribosomes is also expected to result in reduced translation and reduced demand for tRNA use in translation.https://doi.org/10.1371/journal.pgen.1011146.g009individual S. cerevisiae rplΔ or rpsΔ mutant strains, and significant differences between the properties of rplΔ and rpsΔ mutant strains [24,25,50].Presumably the same is true in S. pombe rplΔ or rpsΔ mutant strains.We note in this connection that different tested S. pombe rplΔ or rpsΔ mutations have different growth properties in 3-AT.
According to the proposed model, the mechanism by which rpl or rps mutations suppress will depend on which specific consequence of reduced ribosome concentration is more important in the particular modification mutant.In S. pombe trm8Δ mutants, the GAAC activation is large and concomitant with the onset of a growth defect and the initial loss of tRNA, and the rpl mutations unambiguously suppress the GAAC activation, whereas in S. pombe tan1Δ mutants, the GAAC activation is not as large, and suppression occurs more because of the reduced need for tRNA.As both the reduced GAAC activation and the reduced need for tRNA act in the same direction, to suppress the growth defect in S. pombe, the dominant effect of the rpl mutation (inhibition of GAAC activation or reduced need for tRNA) will depend on the modification mutant, and on the specific conditions of the growth experiment.The situation is more complicated in S. cerevisiae trm8Δ trm4Δ mutants, because the mutations in ribosomal protein genes are expected to exert their effects in two opposite directions.Thus, in S. cerevisiae trm8Δ trm4Δ mutants, rplΔ and rpsΔ mutations are expected to exacerbate the growth defect by preventing GAAC activation, resulting in increased loss of tRNA (Fig 9 ), and are also expected to suppress the growth defect by reducing tRNA demand.The result of these two opposite effects is a more modest predicted effect of rplΔ and rpsΔ mutations on the growth of S. cerevisiae trm8Δ trm4Δ mutants than on the growth of S. pombe trm8Δ or tan1Δ mutants.
Although we do not know why S. pombe trm8Δ and tan1Δ mutants differ in their GAAC activation and its suppression by rplΔ mutants, we speculate that the different effects of modification mutants could arise from known complexities in GAAC activation or the different structures of collided ribosomes.In S. cerevisiae, both the ribosome quality control (RQC) pathway and the GAAC pathway respond to collided ribosomes, and are known to compete with some treatments, whereas other treatments that result in collided ribosomes induce RQC, but not GAAC responses [26].Furthermore, collided ribosomes are known to have different architectures, which could in turn differentially affect induction of the GAAC and/or RQC responses [55].Thus, a particular modification mutant could in principle lead to ribosome collisions that differentially favor the GAAC pathway compared to the RQC pathway.
Based on our model, it seems likely that mutations in rpl or rps genes will similarly suppress the growth defect of other S. pombe modification mutants, particularly those in which the growth defect is due to RTD of an elongator tRNA.Moreover, as the connection between the ribosome and body modification mutants extends to S. cerevisiae trm8Δ trm4Δ mutants, albeit with different phenotypic consequences, it seems likely that similar modification mutants in other organisms, including humans, will be affected as in S. pombe or S. cerevisiae by mutations or conditions that reduce the number of translating ribosomes.

Yeast media and growth conditions
S. pombe strains were grown in rich (YES) media (containing 0.5% yeast extract, 3% glucose, and supplements of 225 mg/l of adenine, uracil, leucine, histidine and lysine), or Edinburgh minimal media complete (EMM-C) containing glucose and 225 mg/l of all amino acids, adenine, and uracil, as well as 100 mg/l of para-amino benzoic acid and inositol, and 1125 mg/l of leucine for Leu -auxotrophs [51].For temperature shift experiments, cells were grown in YES or EMMC-Leu media at 30˚C to OD 600 ~0.5, diluted to ~0.1 OD in pre-warmed media and grown for 9-10 hours as indicated [23].

Bulk RNA preparation and northern blot analysis
For northern analysis, 3 biological replicates were grown in parallel, and then bulk RNA was isolated from pellets derived from 2 ml culture using acid washed glass beads and phenol, separated on a 10% polyacrylamide (19:1), 7M urea, 1X TBE gel, transferred to Amersham Hybond-N + membrane, and analyzed by hybridization to 5' 32 P-labeled DNA probes (S5 Table ) as described ( [13].

Isolation and purification of tRNA, and analysis of nucleoside modifications
S. pombe tan1Δ mutants and WT cells were grown to OD ~1.0 in YES media at 30˚C.Then bulk RNA was extracted from ~300 OD of pellets with hot phenol, and tRNA Leu(CAA) was purified using the 5'-biotinylated oligonucleotide ONV22 (5' TGGTGACCAGTGAGGGAT TCGAAC, complementary to residues 76-53, using standard nomenclature), digested to nucleosides, and analyzed by HPLC as previously described [59].

Quantitative RT-PCR analysis
Strains were grown in triplicate to log phase and bulk RNA was prepared from 2-5 OD pellets using acid washed glass beads and phenol, treated with RQ1 RNase-free DNase (Promega), reverse transcribed with Superscript II Reverse Transcriptase, and cDNA was analyzed by quantitative PCR as described [60].

Polysome profiling
Polysome profiling was done essentially as described [25], except that cells were supplemented with cycloheximide immediately before harvest [42].Strains were grown in liquid YES media at 30˚C to mid-log phase, diluted to OD 600 ~0.1, and grown to OD 600 ~0.5, supplemented with cycloheximide to 100 μg/ml, and then 120 OD-ml of cells were immediately poured onto 200 g crushed ice, and cells were centrifuged, washed, and quick frozen on dry ice.Cell lysates were made using glass beads in Polysomal Lysis Buffer (20 mM Tris-HCL pH 7.5, measured at 20˚C; 50 mM KCl; 4 mM MgCl 2 ; 1 mM DTT; 100 μg/ml cycloheximide, and 1 mM PMSF) containing one tablet Roche protease inhibitor table, 0.2 mg/ml Heparin, and 120 units/ml RNasin, and then 200 μl of cell lysate was layered on the top of a 10-50% (w/v) gradient of sucrose concentration made in Polysomal Lysis Buffer, and samples were centrifuged using an SW-41 rotor (Beckman Coulter) for 3 hours at 35,000 rpm at 4˚C.Gradients were fractionated using a BioComp Gradient Station and analyzed by absorbance at 254 nm.

Spontaneous suppressors and whole-genome sequencing
To select spontaneous suppressors of S. pombe trm8Δ or tan1Δ mutants, individual colonies were inoculated into YES media at 30˚C and ~10 7 cells of overnight cultures were plated on YES media at 38˚C and 39˚C.Whole genome sequencing was performed by the University of Rochester Genomics Center or the Cornell Genomics Facility at a read depth of 20-80 per genome nucleotide.

Fig 1 .
Fig 1.A class of S. pombe trm8Δ suppressors have mutations in rpl genes, minimally restore tRNA levels, and prevent activation of the GAAC pathway.(A).S. pombe trm8Δ suppressors with distinctive growth properties have mutations in rpl genes.S. pombe trm8Δ mutants and suppressors as indicated were grown overnight in YES media at 30˚C, diluted to OD 600 ~0.5, serially diluted 10-fold in YES media, and 2 μl was spotted on plates containing YES, EMMC-His, EMMC-His with 10 mM 3-AT, and YES with 5-FU (30 μg/ml), as indicated.(B).The trm8Δ

Fig 4 .
Fig 4. S. pombe tan1Δ mutants lack 4-acetylcytidine and are temperature sensitive due to decay of tRNA Leu(AAG)and tRNA Leu(UAG) .(A). S. pombe tan1Δ mutants lack ac 4 C in their tRNA Leu(CAA) .S. pombe WT and tan1Δ mutants were grown in triplicate in YES media at 30˚C to late log phase, and tRNA Leu(CAA) was purified from bulk RNA, digested to nucleosides, and analyzed for modifications by HPLC as described in Materials and Methods.(B).S. pombe tan1Δ mutants are temperature sensitive.S. pombe WT and tan1Δ mutants were transformed with a [leu2 + ]

Fig 5 .Fig 6 .
Fig 5.The temperature sensitivity of S. pombe tan1Δ mutants is efficiently suppressed by rplΔ and rpsΔ mutations.(A).A reconstructed tan1Δ rpl2802Δ mutant suppresses the S. pombe tan1Δ growth defect.The rpl2802Δ mutation was introduced into WT, and tan1Δ strains, and then WT, tan1Δ, rpl2802Δ, and tan1Δ rpl2802Δ strains were grown overnight in YES media and analyzed for growth on media containing YES, EMMC-His, EMMC-His with 10 mM 3-AT, and YES with 5-FU (30 μg/ml), as indicated.(B).A reconstructed tan1Δ rpl1701Δ mutant suppresses the S. pombe tan1Δ growth defect, as does each of two other tested deletions of ribosomal proteins.The rpl1701Δ, rps23Δ, and rpl502Δ mutations were introduced into WT and tan1Δ strains, and then strains were grown overnight in YES media and analyzed for growth on media containing YES, EMMC-His, EMMC-His with 10 mM 3-AT, and YES with 5-FU (30 μg/ml), as indicated.https://doi.org/10.1371/journal.pgen.1011146.g005

Fig
Fig 6A was analyzed for GAAC activation and normalized as described in Fig 1D.(D).The temperature sensitivity of a tan1Δ mutant is more efficiently suppressed in tan1Δ rpl502Δ strains than in tan1Δ gcn2Δ strains.WT and tan1Δ strains with rpl502Δ or gcn2Δ mutations were grown overnight in YES media and analyzed for growth on media containing YES or EMMC-His, as indicated.https://doi.org/10.1371/journal.pgen.1011146.g006

Fig 7 .
Fig 7.Each of several S. pombe rplΔ and rpsΔ strains have growth defects, compared to WT strains, and have polysome profiles indicative of reduced translation.(A).An S. pombe rpl502Δ strain grows distinctly slower that a WT strain.An S. pombe rpl502Δ mutant and WT strain were grown in liquid YES media at 30˚C to mid-log phase (OD 600 0.24 and 0.294 respectively), diluted to OD 600 ~0.1 in the same media, and growth was monitored at 30˚C every hour for 8 hours to monitor the growth rate.Data are fit to an exponential, with R 2 values as shown, and generation times determined as ln2/exponent*60 min.(B).Each of several S. pombe rplΔ and rpsΔ strains have growth defects.S. pombe WT, rplΔ, and rpsΔ strains as indicated were grown in triplicate in liquid YES media at 30˚C to mid-log phase (OD 600 range from 0.24 to 0.54), and diluted and grown as described in (A) to derive generation times and R 2 values from exponential fits (S1 Table).(C).S. pombe rpl502Δ, rpl2802Δ, and rps23Δ strains have polysome profiles indicative of reduced 60S or 40S subunits, and reduced translation.Strains as indicated were grown in YES media at 30˚C to mid-log phase, and cell lysates were analyzed by polysome profiling in a 10-50% sucrose gradient, as described in Materials and Methods.Wild-type lysate is compared with lysates from rpl502Δ, rpl2802Δ, and rsp23Δ mutant strains.Peaks corresponding to halfmer polysomes are indicated by arrows.

Fig 8 .
Fig 8. Deletion of any of several ribosomal protein genes exacerbates the temperature sensitivity of S. cerevisiae trm8Δ trm4Δ mutants in YPD Media.rpl genes (A) and rps genes (B) as indicated were deleted in S. cerevisiae WT and trm8Δ trm4Δ strains, and cultures were grown and analyzed on the indicated media.https://doi.org/10.1371/journal.pgen.1011146.g008

Fig 9 .
Fig 9.A model describing the roles of rpl, rps mutations in the biology of S. pombe trm8Δ and S. cerevisiae trm8Δ trm4Δ mutants.In both S. pombe trm8Δ mutants (red) and S. cerevisiae trm8Δ trm4Δ mutants (green), elevated temperature (38.5˚C and 32˚C respectively) triggers RTD of the corresponding tRNA (decay of tRNA Tyr(GUA) by Dhp1/Rat1 in S. pombe and decay of tRNA Val(AAC) by Rat1/Dhp or Xrn1 in S. cerevisiae).In each organism, the reduced levels of the corresponding tRNA result in increased ribosome collisions and activation of the GAAC pathway (blue), leading to further loss of tRNA Tyr(GUA) in S. pombe and prevention of further loss of tRNAVal(AAC) in S. cerevisiae[23].In both organisms, the rpl and rps mutations (brown) result in a reduced concentration of ribosomes, leading to inhibition of ribosome collisions and reduced GAAC activation in S. pombe, and presumed to reduce GAAC activation in S. cerevisiae.In both organisms, the reduced concentration of ribosomes is also expected to result in reduced translation and reduced demand for tRNA use in translation.