Pseudouridine-mediated translation control of mRNA by methionine aminoacyl tRNA synthetase

Abstract Modification of nucleotides within an mRNA emerges as a key path for gene expression regulation. Pseudouridine is one of the most common RNA modifications; however, only a few mRNA modifiers have been identified to date, and no one mRNA pseudouridine reader is known. Here, we applied a novel genome-wide approach to identify mRNA regions that are bound by yeast methionine aminoacyl tRNAMet synthetase (MetRS). We found a clear enrichment to regions that were previously described to contain pseudouridine (Ψ). Follow-up in vitro and in vivo analyses on a prime target (position 1074 within YEF3 mRNA) demonstrated the importance of pseudouridine for MetRS binding. Furthermore, polysomal and protein analyses revealed that Ψ1074 mediates translation. Modification of this site occurs presumably by Pus6, a pseudouridine synthetase known to modify MetRS cognate tRNA. Consistently, the deletion of Pus6 leads to a decrease in MetRS association with both tRNAMet and YEF3 mRNA. Furthermore, while global protein synthesis decreases in pus6Δ, translation of YEF3 increases. Together, our data imply that Pus6 ‘writes’ modifications on tRNA and mRNA, and both types of RNAs are ‘read’ by MetRS for translation regulation purposes. This represents a novel integrated path for writing and reading modifications on both tRNA and mRNA, which may lead to coordination between global and gene-specific translational responses.


INTRODUCTION
Over the past decades, >100 post-transcriptional RNA modifications have been identified in all kingdoms of life (1). One of the most common RNA modifications in living cells is pseudouridine ( ) (2,3). The isomerization of uridine to pseudouridine has a similar base pairing as uridine, yet it enables stabilization of RNA structure by an extra hydrogen bond (4,5). In yeast, non-coding RNAs such as rRNA, snoRNA, snRNA and tRNA are highly enriched with pseudouridine at dozens of specific positions (6,7).
The isomerization of pseudouridine is catalyzed by two groups of enzymes ('writers'). The first group, operating as an RNP complex, uses guide-RNA-based pseudouridine isomerization in which the essential Cbf5 enzyme is guided by H/ACA box snoRNA to its target rRNA sites (7)(8)(9). The second pseudouridine synthase (PUS) group is the standalone enzyme family that catalyzes the reaction without additional proteins or RNAs (10). Pseudouridine isomerization of cytosolic and mitochondrial tRNAs (11)(12)(13)(14)(15)(16) is catalyzed by these enzymes.
Recently, high-throughput pseudouridine-seq studies revealed hundreds of pseudouridine sites in 509 different mR-NAs. Most of this mRNA pseudouridylation activity was catalyzed by the stand-alone tRNA PUS family. Moreover, those studies indicated that pseudouridylation may be reg-ulated in response to environmental changes (17)(18)(19). While in vitro analysis revealed an impact on translation (20), the in vivo impact of this post-transcriptional modification on mRNA activity remains largely unknown.
Modifications in RNA molecules usually serve as recognition sites to proteins ('readers') that exert a regulatory function. To date, the only identified pseudouridine reader is the yeast Prp5 RNA helicase. Prp5 interacts with 42 and 44 within U2 snRNA and promotes pre-mRNA splicing (21). Surprisingly, no mRNA or tRNA pseudouridine readers have been identified thus far.
The family of aminoacyl tRNA synthetases (aaRS) sustains extensive interactions with pseudouridylated RNAs (i.e. tRNAs). aaRSs identify their cognate tRNA and catalyze the charging reaction of a specific amino acid at the tRNA 3 end (22). The interaction of a specific tRNA with its cognate aaRS depends on unique tRNA identity elements. These elements are located mainly at the acceptor stem and the anticodon stem-loop (23)(24)(25). tRNA identity elements are also a hotspot of RNA modifications. These post-transcriptional RNA modifications can change the tRNA structure and may play a central role in tRNA-aaRS interactions (26). However, the in vivo impact of tRNA pseudouridylation on aaRS binding is not known. The association of aaRSs with RNA is not restricted to tRNAs (27,28). Recent interactome studies revealed that aaRSs interact with polyA RNA, likely mRNAs (29)(30)(31). Various anticodon-like elements within mRNA appeared key for aaRS-mRNA interaction and regulation (32)(33)(34); however, a role for mRNA modification is unknown.
Herein, we performed an in vivo binding-site mapping for Saccharomyces cerevisiae MetRS. Comparing these data to previously published pseudouridine-seq data revealed that MetRS preferentially binds -enriched regions. We revealed that at position 1074 of YEF3 mRNA and 31 of elongator tRNA Met are important for MetRS binding. The former interaction regulates YEF3 protein synthesis while the latter has a global effect on translation. Furthermore, our data suggest that both modifications are made by Pus6, suggesting a pivotal role for this enzyme in regulating global and gene-specific translation.

Fragmentation of RNA and ImmunoPrecipitation (fRIP)
MetRS-TAP strain was grown in YPD to mid-logarithmic phase and subjected to cross-linking by the addition of formaldehyde (0.1% final concentration) for 10 min at room temperature. Cross-linking was terminated with 0.125 M glycine for 3 min, and cells were lysed in Buffer B (

RNA-binding proteins ImmunoPrecipitation and qPCR (RIP-qPCR)
RIP-qPCR analysis was done as described in (33). Briefly, for mRNA RT-qPCR, RNA was reverse-transcribed using a high-capacity cDNA reverse transcription kit (Maxima) according to the manufacturer's instructions. Gene-specific levels were determined in a 25 l reaction volume in triplicate with a Power SYBR Green PCR Master Mix ® (Applied Biosystems) two-step RT-PCR method following the manufacturer's instructions using primers for the indicated genes (Supplementary Table S1). All qPCR reactions used the following parameters: 50 • C for 2 min, 95 • C for 10 min, and then 30 s at 95 • C and 1 min at 60 • C for 40 cycles. Results were analyzed with Applied Biosystems 7500 Real-Time PCR Software v2.0.6. Fold change was calculated using either 2 −( Ct) or 2 −( Ct) .

tRNA quantification
tRNA RT-qPCR was done as described in (35). Briefly, tRNA was reverse-transcribed using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). Reaction temperature was elevated to 60 • C, and reversetranscription elongation was extended to 30 min to efficiently overcome tRNA modification and secondary structure. tRNA-specific levels were determined in a 25 l reaction volume in triplicate with a Power SYBR Green PCR Master Mix ® (Applied Biosystems) following the manufacturer's instructions using primers for the indicated tRNA as described in (35). Results were analyzed with Applied Biosystems 7500 Real-Time PCR Softwarev2.0.6. Fold change was calculated using 2 −( Ct) . tRNA RT-qPCR products were cloned into a pGEM system and subjected to Sanger sequencing to validate amplification of the proper tRNA. All primers are listed in Supplementary Table S1.

CRISPR/Cas9 point mutagenesis
Point mutations were introduced using the single-plasmid CRISPR/Cas9 protocol (41,42). Briefly, plasmid (bRA66 backbone (Addgene #100952)) expressing guide RNA recognizing the mutation region was transformed into MetRS-GFP strain together with the double-stranded 80 bp DNA fragment that contains the desired mutation at its center (Supplementary Table S1). Transformants were grown on YP-Gal Hygromycin (200 g/ml) plate to induce Cas9 expression. Three viable clones from the transformation with the 80 bp fragment were isolated and the modification region was sequenced; all included only a single change, T1074C.

S methionine pulse labeling
Cells (10 ml) were grown to mid-log phase in synthetic complete (SCD) medium, washed and resuspended in 10 ml SC without methionine. 35 S methionine (NEG709A005MC, PerkinElmer) was added and growth was continued for another 30 min. The translation was quenched by the addition of cycloheximide (100 g/ml final concentration). Proteins were extracted and resolved on 10% PAGE. Gels were exposed to a phosphorimager and radioactive signals were quantified by ImageJ.

Polysomal analysis
Polysomal analysis was done as described in (43). Briefly, 100 ml cells grown in YPD to mid-logarithmic stage were harvested and lysed by a bead beater in 0.4 ml of lysis buffer (20 mM Tris-HCl at pH 7.4, 140 mM KCl, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol, 100 g/ml cycloheximide, 1 mg/ml heparin, 1% Triton X-100). The lysate was cleared and centrifuged for 15 min at 9000 g at 4 • C and the supernatant was loaded onto a 12 ml 10-50% linear sucrose gradient. Gradients were centrifuged in a SW41 rotor (Beckman) at 35 000 rpm for 160 min and polysomal profiles were determined by monitoring RNA absorbance at 254 nm with ISCO UA-6. Polysomal association was calculated by dividing the polysomal area by the area from the beginning of the 40S to the end of the graph. The entire gradient was fractionated to fractions of similar volume. RNA was extracted from each fraction by the addition of an equal volume of 8 M guanidinium HCl and two volumes of 100% ethanol, overnight incubation at -20 • C and centrifugation at 13 000 rpm for 30 min. Pellets were washed with 80% cold ethanol, and usually half of the sample was subjected to northern blot analysis with the indicated probes.

Fragmented RIP-seq (fRIP-seq) protocol for mapping MetRS-bound RNA regions
Our previous RNA-binding protein ImmunoPrecipitation (RIP-seq) study revealed a subset of mRNAs that are highly associated with yeast methionine aminoacyl synthetase (MetRS) (33). To narrow down RNA regions that MetRS is associated with, we established a fragmentation of RNA and RIP-seq (fRIP-seq) protocol. In fRIP-seq, cellular lysate of a TAP-tagged MetRS strain (or, for that matter, any other RNA-binding protein) is subjected to a mild RNase I treatment prior to RNA-binding protein immunopurification (RIP) with IgG Sepharose beads. Fragmented samples from either before RIP (Input) or those eluted from the beads by TEV cleavage (IP) were analyzed by next-generation sequencing ( Figure 1A). The genomic position of Input and IP fragments was compared and en-riched positions ('peaks') were defined. While we applied this approach herein to MetRS-TAP, we note that we successfully applied it to GFP-tagged proteins using Chromotek GFP-Trap beads (Levi O. et al., in preparation).
The mild fragmentation in fRIP generates RNA fragments of ∼200 nts, thereby defining the resolution of the approach (Supplementary Figure S1A). About 25% of tagged MetRS proteins were detected in the elution sample, in- Venn diagram was generated from peaks identified as bound by the MetRS fRIP-seq analyses and specific mRNA ribonucleotide that was found to contain pseudouridine by -mapping studies (17)(18)(19). P-value was calculated by the Hypergeometric test. (B) MetRS bound peaks were separated into those that do not contain (non-) and those that do. Box plot presents the fold enrichment in binding compared to untagged control. P -value was calculated by the two-tailed t-test. (C) GO terms processes analysis was done to -containing peaks genes that are bound by MetRS using SGD GO Term Finder Version 0.86. dicating the efficiency of MetRS isolation (Supplementary Figure S1B and C). Western analysis for Hxk1 control protein did not detect any signal at the bound sample, indicating the specificity of isolation. Furthermore, silver staining of all proteins in the bound sample detected only marginal levels of proteins other than MetRS, and all were detected in the control purification of an untagged strain (Supplementary Figure S1C). Overall, under these conditions, MetRS appears to be efficiently and specifically isolated.
RNA samples from two biological repeats of fragmented cell lysate (Input) and the IP sample were sequenced. A similar scheme was applied to the parental untagged strain to control for non-specific binding. A high correlation is apparent between the Input signals of the two MetRS repeats and the untagged strain ( Figure 1B). Thus, the TAP-tagging of MetRS does not lead to a general effect on gene expression. Analysis of the IP samples revealed that while both biological repeats of MetRS-TAP were very similar (Pearson correlation of 0.907), both were significantly different from the untagged control IP ( Figure 1C). Thus, the fRIPseq protocol is very reproducible, attested by the high correlation between biological repeats, and highly specific, apparent from the low correlation with the untagged strain. Figure 1D presents read maps of four representative mR-NAs. The high efficiency and specificity of binding is apparent upon comparison to the Input sample and/or the Untagged strain data.
Applying a peak calling algorithm (40) to define mRNA regions that are significantly higher compared to the overall background of the mRNA and reproducible in both biological repeats, identified 145 such regions scattered within 131 mRNAs ( Figure 1E, Supplementary Table S2). Interestingly, these regions are mostly localized within the coding region of the mRNA ( Figure 1F). DREME sequence alignment (44) identified several significant motifs, all containing UA dinucleotide, with GNUANC being the most significant ( Figure 1G, Supplementary Figure S1D); we did not find any resemblance to sequences within canonical targets of MetRS (i.e. tRNA Met ) (45). GO term analysis revealed that binding is significantly enriched to mRNAs encoding translation factors (GO: 0002181) ( Figure 1H, Supplementary Table S3). Altogether, these data suggest that MetRS regulates expression of translation factors through binding to their coding regions.

MetRS is a pseudouridine reader
The canonical function of MetRS is charging of initiator and elongator tRNA Met with methionine. These tRNAs are known to be modified extensively, in particular at uridines that are modified into pseudouridine ( ). Therefore, we examined whether MetRS-bound RNAs are enriched with . Three independent pseudouridine-seq studies were previously published (17)(18)(19). While these datasets are likely incomplete, as apparent from the low overlap in their hits (46,47), altogether they report on 750 sites of within S. cerevisiae transcriptome. We compared these sites with the MetRS-bound fragments detected here, and a significant number (49 of 145, P-value 1.8e −48 ) appeared to overlap (Figure 2A, Supplementary Table S4). Moreover, the peaks that contain a appear to be better bound by MetRS compared to the 96 not reported to be pseudouridylated ( Figure  2B). Thus, pseudouridine enhances MetRS binding to its target regions. GO term analysis for the 49 peaks that contain pseudouridine recaptured the enrichment to mRNAsencoding cytosolic translation factors ( Figure 2C, Supplementary Table S3), suggesting that regulation of this family occurs through .

of YEF3 is important for MetRS binding
MetRS appeared to significantly bind a region around nt 1000 of the mRNA-encoding eEF3 elongation factor (YEF3) ( Figure 3A, red bar). This region was reported by two independent studies to include a site at position 1074 downstream to the start codon (chrXII:637854) (18,19). Nucleotide 1074 is a wobble position of a UUU codon (Phe amino acid). Position 1075 is occupied by an A, thus generating a UA dinucleotide, in line with the core apparent by computational predictions (Supplementary Figure  S1D). To examine the possibility that MetRS binds this site, we expressed S. cerevisiae MetRS protein in bacteria and purified it through MBP column ( Figure 3B). Purified protein was mixed and crosslinked with labeled 9-mer oligos identical to the 1074 region with either or U at its center ( Figure 3C). A 2-fold higher association was observed in the -containing 9-mer compared to the mutant ( Figure  3D). Importantly, the addition of a competitor cold tRNA abolished binding to either oligo ( Figure 3E). We note that attempts to derive exact disassociation constant (K D ) using Red bar under the map indicates the MetRS binding region, as identified by the MACS2 callpeak (40). Arrows point to positions revealed by (18,19) along YEF3 ORF. HS indicates sites apparent upon heat shock. (B) Coomassie blue staining of MetRS purification steps. Noninduced lane (NI) contains a protein sample from bacteria transformed with MBP-MetRS plasmid before induction. Input (I) contains a protein sample from bacteria after expression induction. Elution includes the protein sample eluted from the MBP column. (C) Labeled 9-mer oligonucleotide (identical to the sequence surrounding 1074), either with or without at its center, was UV-crosslinked to the indicated extent with purified MetRS and resolved on PAGE. Gel was dried and exposed to a phosphorimager. (D) Quantification of the MetRS-bound oligo to free oligo ratio from two independent crosslinking experiments. ** indicates P -value ≤0.01 for the difference between and U containing oligos at each UV exposure. Note that error bars for the U oligo are smaller than the data points. (E) UV crosslinking was done at the indicted amounts. Competitor cold yeast tRNA was added prior to crosslinking where indicated. (F) Point mutation was introduced by CRISPR/Cas9 into the genomic site of YEF3, replacing the encoded U1074 with C. Steady-state levels of wildtype (WT) and mutant (U1074C) YEF3 mRNA were quantified by RT-qPCR analysis from three independent biological repeats, each with three technical repeats, normalized to ACT1 mRNA levels. (G) MetRS-GFP expressing yeast strains, co-expressing endogenously either WT or U1074C YEF3 mRNA, were subjected to RIP. Amounts of bound YEF3 mRNAs were quantified by RT-qPCR analysis and normalized to their expression levels. The histogram presents the quantification of two independent biological repeats, each with three technical repeats. P-value was calculated by the dependent samples one-tailed t-test.
EMSA were futile, probably because affinity to the 9-mer was too low (not shown). Thus, the optimal MetRS binding site is probably longer.
To examine the significance of at position 1074 ( 1074) in vivo, we replaced the thymidine (T) in this genomic position with cytosine (C), thereby eliminating pseudouridylation, yet maintaining the same amino acid (both UUU and UUC encode Phe). The mutation was introduced in a seamless manner by CRISPR/Cas9, without the insertion of any selection marker or any other change in the YEF3 gene (42). Analysis of steady-state YEF3 mRNA levels revealed a small and insignificant change in the U1074C mRNA levels ( Figure 3F). Significantly, however, RIP followed by RT-qPCR to cells expressing MetRS-GFP with  Figure S3). Arrowheads point to the fraction with the highest signal ('peak fraction') per mRNA. Error bars are SEM of three independent biological repeats. Note that the fraction collector was inadvertently set to collect 10 fractions for U1074 gradients while the WT were separated into 11 fractions. Yet in both cases the entire gradient was collected (gradients' OD254 profiles are provided in Supplementary Figure S3A). either normal or U1074C YEF3 mRNA revealed a strong reduction in MetRS association with U1074C mutant compared to the WT transcript ( Figure 3G). This demonstrates that 1074 is important for MetRS association with YEF3 mRNA in vivo.

confers a translation regulatory role
Protein binding to an mRNA usually serves expression regulation purposes, which can be exerted through changes in mRNA stability or translation. Since we did not detect significant changes in mRNA steady-state levels upon U1074C change ( Figure 3F), we tested possible impact on protein synthesis. Western analysis revealed a clear increase in Yef3 protein levels upon U1074C point mutation ( Figure 4A and B). We note that U1074C mutation leads to a silent codon change from UUU to UUC (both code for Phe), with UUC having a lower codon usage than UUU (18.4 versus 26.1) (48). Since lower usage is associated with lower translation (49), it is unlikely to explain the increased protein amounts we observe. Furthermore, both codons are recognized by the same tRNA (tRNA GAA ) as no tRNA AAA exist S. cere-visiae (45); hence, differences in cognate tRNA abundance also cannot explain the increased protein levels.
To substantiate the link to translation, polysome profiling through sucrose gradients was done to both strains. While ribosomal association of unmutated transcripts (i.e. HXK1 and CDC33) appeared not to change, and possibly even decreased in the U1074C strain, the ribosomal association of the U1074C transcript appeared to increase (i.e. shifted to the right) ( Figure 4C, Supplementary Figure S3B). Thus, 1074 mediates YEF3 translation.

Pus6 affects MetRS binding to tRNA
Pseudouridine synthetases (PUSs) isomerize U to among many tRNAs. The Pus6 member of this family has only one cytosolic tRNA target, elongator tRNA Met . Furthermore, U31 within elongator tRNA Met is the only known cytosolic target of Pus6 (14). We performed RIP analysis for Pus6 tagged by GFP to examine its tRNA association in vivo.
Purification efficiency of this protein appears similar to the other GFP-tagged MetRS strains (Supplementary Figure  S2). RT-qPCR designed for tRNA detection (35) was used to explore its targets. A strong association with tRNA Met elongator is apparent, more than 10-fold higher than the control, non-target tRNA Pro ( Figure 5A). A similar preference was observed for MetRS binding to tRNA Met ( Figure  5B). These data provide an in vivo support to tRNA binding specificity of Pus6 and MetRS. We next wished to determine the impact of pseudouridylation by Pus6. The deletion of Pus6 did not have a significant effect on steady-state levels of either tRNA Met or tRNA Pro , suggesting that the modification is not important for tRNA stability ( Figure 5C). Importantly, however, RIP analysis to MetRS-GFP detected a strong reduction in MetRS binding to elongator tRNA Met in pus6Δ, and a smaller and insignificant impact on binding to tRNA Pro ( Figure 5D). This is not due to a decrease in MetRS purification upon Pus6 deletion (Supplementary Figure S2). To the best of our knowledge, this is the first time that pseudouridylation of a tRNA is linked to tRNA synthetase binding in vivo.

Pus6 mediates global protein synthesis
The impact of pseudouridylation on tRNA Met binding to MetRS may lead to a broad impact on protein synthesis. To examine this, we performed pulse labeling for all cellular proteins by 35 S-met labeling of pus6Δ and its parental strain (WT) (Figure 6). A clear reduction in their synthesis rates upon Pus6 deletion is observed ( Figure 6A and B). Intriguingly, polysomal RNA analysis revealed an increase in OD 254 signal of polysomal fractions in pus6Δ cells ( Figure 6C and D). This suggests that the lower protein levels are due to slowed translation elongation, leading to accumulation of ribosomes on mRNAs upon Pus6 deletion.

Pus6 binds and affects YEF3 translation
Pus6 was previously reported to modify mRNAs (18). Therefore, we tested its possible role in mRNA regulation. In vivo binding of Pus6 to mRNAs was assayed by RIP followed by RT-qPCR ( Figure 7A). A significantly higher association with YEF3 than with an mRNA that was not found to be pseudouridylated (ACT1) was observed. A strong association was also observed with CDC33 mRNA (encoding the translation initiation factor eIF4E), supplementing previous in vitro pseudouridylation analysis by purified Pus6 on CDC33 mRNA (18). Furthermore, Pus6 deletion did not affect steady-state YEF3 mRNA levels ( Figure 7B), yet led to a significant reduction in MetRS association with YEF3 ( Figure 7C). Thus, Pus6 is a 'writer' of pseudouridine on YEF3 that is 'read' by MetRS.
Concomitant with the decreased MetRS association, analysis of YEF3 protein levels revealed a clear increase upon Pus6 deletion ( Figure 7D and E). A similar increase is observed for CDC33, another target of Pus6 ( Figure 7D and E). Polysomal analysis to pinpoint the impact to translation revealed no change in the sedimentation of the control HXK1 mRNA, while both modified mRNAs, CDC33 and YEF3, have a higher polysomal association in pus6Δ cells ( Figure 7F). Together, these data reveal that Pus6 deletion imposes an activating effect on translation of its target mRNAs, presumably through reduced MetRS binding to sites.

DISCUSSION
Functions of aaRS other than tRNA charging are well known and include diverse roles in mRNA regulation (50,51). Interestingly, some of these functions are mediated through recognition of mRNA elements that resemble tRNA elements (24). For example, elements that are similar to anticodons of tRNA His or tRNA Thr were found within mRNA targets of HisRS and ThrRS, respectively, (33,34) and both exert a translation regulatory role. Here, we show for the first time, the involvement of another established tRNA feature, modification, in mRNA recognition by aaRS followed by translation control. Specifically, we show that MetRS interacts with both tRNA and mRNA through a pseudouridine that is generated by Pus6. Interaction with these two types of RNA offers coordination between tRNA charging and mRNA translation. Our working model (Figure 8) poses that Pus6 modifies elongator tRNA Met , and this modification is important for MetRS binding ( Figure 5D). The decreased association with MetRS likely leads to a decrease in tRNA Met charging. We speculate that lower levels of charged tRNA Met reduce rates of translation elongation, hence, the increase in polysomal complexes observed in pus6Δ cells ( Figure 6D). In parallel, Pus6 is a pseudouridine writer on specific mRNA targets ( Figure 7A) (17,18). These sites are probably important for MetRS binding to mRNA, therefore upon Pus6 deletion the association of MetRS decreases ( Figure 7C). Importantly, we show that Pus6 deletion leads to increased translation of mRNAs that are known to have a pseudouridine ( Figure 7D-F). Altogether, we propose that pseudouridines introduced by Pus6 are read by MetRS to exert a global and a gene-specific translation regulation.
What is the regulatory purpose of this pathway? The opposing effects on general and gene-specific translation echo the well-studied yeast amino acid starvation pathway (52). In response to amino acids depletion, GCN2 phosphorylates eIF2␣, which leads to sequestration of eIF2B, followed by inhibition of global translation yet increased translation of GCN4 mRNA (53,54). While the mechanistic details of our observation are different, it suggests another path for coordination between global and gene-specific regulation in response to environmental changes. Thus far, we were unable to identify the environmental conditions upon which Pus6 is activated; the obvious predictions of methionine or general amino acid changes did not appear to change Pus6 levels (data not shown). Nevertheless, pseudouridylation by Pus6 appears as a new means to coordinate between tRNA and mRNA functions and global and gene-specific translation regulation. Since pseudouridine is among the most abundant modifications in RNA, we expect that regulation will be implicated on many transcripts. Furthermore, a significant fraction of mRNAs that include and bound by MetRS encode components of the translation machinery. These proteins are key regulators of many cellular processes, suggesting an even broader impact on the expression.
The binding of MetRS within the coding region suggests that binding inhibits ribosome transit along the mRNA, thereby affecting the elongation rate. In such a case, one would expect to see an increase in ribosome density upstream to the MetRS binding site. However, a previous Ri- Figure 7. Pus6 regulates translation of pseudouridylated mRNAs. (A) GFP-tagged Pus6 cells were subjected to RIP followed by RT-qPCR to the indicated mRNAs. The histogram presents the quantification of two independent biological repeats. P-values are relative to ACT1 mRNA levels and were calculated by the dependent samples one-tailed t-test. (B) RT-qPCR analysis of YEF3 mRNA levels normalized to ACT1 in WT and pus6 deleted cells (pus6Δ). The histogram presents the quantification of two independent biological repeats. (C) MetRS-GFP expressing cells either containing (WT) or deleted of Pus6 (pus6Δ) were subjected to RIP followed by RT-qPCR to YEF3 mRNA. The histogram presents the quantification of two independent biological repeats, normalized to the signal in the Input sample. P-value was calculated by the dependent samples one-tailed t-test. (D) Western analysis for the indicated proteins in WT and in pus6Δ strains. Samples from two independent protein preparations are presented. Signals are from the same blot, from which irrelevant lanes were cropped. (E) Quantification of Yef3p and Cdc33p signals in WT and pus6Δ cells normalized to the Hxk1p signal. Results are from three independent biological repeats. Error bars are SEM, and P-value was calculated by the dependent samples one-tailed t-test. (F) WT and pus6Δ cells were subjected to polysomal separation followed by northern analysis for YEF3, CDC33 and HXK1 transcripts. Graphs present the quantification of northern signals within all polysomal fractions. Arrowheads point to the fraction with the highest signal ('peak fraction') per mRNA. Results are from three independent biological repeats, error bars are SEM.  .g. YEF3). These modifications enhance MetRS binding on both types of RNA. MetRS binding to tRNA Met presumably affects its charging and thereby has a global impact on translation, while MetRS binding to an mRNA leads to a gene-specific impact.
bosome Density Mapping along YEF3 mRNA from cells grown under similar conditions did not reveal a higher density of ribosomes at that region compared to other regions along the mRNA (55). This suggests that impact is not through an elongation arrest. Due to the lack of apparent impact on elongation, we speculate that under normal conditions, the pseudouridine at position 1074 affects the translation initiation rate. How a modification that is located in the midst of an ORF affects events at the 5 end of the transcript is yet to be revealed.
Yef3 protein is an elongation factor that assists the ribosomes' movement along the ORF and plays a role in ribosome recycling (56). It may also play a non-translational role (57). Recently, the impact of Yef3p depletion on ribosome density was studied at high resolution (58). Interestingly, the depletion of Yef3p led to a higher representation of methionine codons at the P-site of stalled ribosomes. This implies a preferred role for Yef3p in assisting ribosomes stalled on these codons. Reduction in charged tRNA Met levels may lead to this situation. Therefore, we speculate that under some conditions, the activity of Pus6 is repressed, leading to lower charging of tRNA Met and ribosomes' stalling. This condition will entail an increased need for Yef3p to rescue stalled ribosomes. This reiterates the tight interconnections between gene-specific and global regulation mediated by Pus6 and MetRS.