A stress‐induced tyrosine‐tRNA depletion response mediates codon‐based translational repression and growth suppression
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (Review Commons)
Abstract
Article activity feed
-
-
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
Reviewer #1 (Evidence, reproducibility and clarity (Required)):
The manuscript by Huh et al. reports that oxidative stress causes fragmentation of a specific tyrosine pre-tRNA, leading to two parallel outcomes. First, the fragmentation depletes the mature tRNA, causing translational repression of genes that are disproportionally rich in tyrosine codon. These genes are enriched for those involved in electron transport chain, cell cycle and growth. Second, the fragmentation generates tRNA fragments (tRFs) that bind to two known RNA binding proteins. Finally, the authors identify a nuclease that is needed for efficient formation of tyrosine tRFs.
Comment 1: …
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
Reviewer #1 (Evidence, reproducibility and clarity (Required)):
The manuscript by Huh et al. reports that oxidative stress causes fragmentation of a specific tyrosine pre-tRNA, leading to two parallel outcomes. First, the fragmentation depletes the mature tRNA, causing translational repression of genes that are disproportionally rich in tyrosine codon. These genes are enriched for those involved in electron transport chain, cell cycle and growth. Second, the fragmentation generates tRNA fragments (tRFs) that bind to two known RNA binding proteins. Finally, the authors identify a nuclease that is needed for efficient formation of tyrosine tRFs.
Comment 1: The authors should include a short diagram indicating the various known steps of pre-tRNA fragmentation (perhaps as a supplement) for general readers.
Response: We thank the reviewer for their suggestion. Pre-tRNA fragmentation is still an unknown field but an initial introduction is best seen from pre-tRNA processing where there is a cleavage event for pre-tRNAs with an intron. This is a complex subject but a recent review from Hopper and Nostramo has done an excellent job in in describing the current field in yeast and vertebrate species (Hopper and Nostramo, Front. Genet., 2019). We have added this citation and new text in the manuscript about pre-tRNA processing for general readers to follow up on. We feel that a supplementary figure might be a bit too brief in describing the knowns and unknowns of pre-tRNA processing and fragmentation.
Comment 2: I find the enrichment for mitochondrial electron transport chain (ETC) curious. The ETC includes several oxidoreductases, which may be rich in tyrosine as it is a common amino acid used in electron transfer. The depletion of the tyrosine tRNA from among many tRNAs under oxidative stress may not be incidental but related to an attempt by the cell to decrease oxygen consumption to avoid further oxidative damage. The authors could further mine their data to corroborate this hypothesis. For example, are the ETC genes among the targets of the RNA binding proteins targeted by tyrosine tRFs? This could potentially connect the effects of mature tRNA depletion and tRFs.
Response: We thank the reviewer for this very interesting comment and insight, which had not occurred to us. The relationship between this response and oxidoreductase regulation could be a factor in both the tRNA and tRF modulations seen in our cells. Interestingly, we find that many oxidoreductases genes (such as the NDUF family) are bound by hnRNPA1 by CLIP. In new data, we have done stability experiments with the tRF (new Fig 7E-F) to show the regulon of hnRNPA1 is modulated with overexpression and LNA against the tRF, revealing that this tRNA fragmentation response modulates expression of certain oxidoreductase genes. However, we do not see clear and significant differences for ETC genes in particular. As hnRNPA1 is known to act as both a promoter and destabilizer of genes depending on context, it is likely that further and more detailed work will be needed to parse this hypothesis out in future studies.
__Comment 3: __In figure 4A, the authors should provide the tyrosine codon content of the overlap genes and show how much it differs from a randomly selected sample.
__Response: __We have identified an error in our manuscript where the overlap actually identifies 109 proteins rather than the 102 reported in the original manuscript. We apologize for this oversight. As for the overlap proteins, we plotted the downstream proteins detected in the proteome by mass spectrometry based off on Tyr-codon content. As explained in the text, the targets we tested were chosen for having higher than median levels of Tyr-codon, as seen in the histogram, and for showing some of the greatest reduction after Tyr tRNA-GUA depletion (Fig S4A). The other proteins found in the overlap will fall in a similar pattern along the histogram.
Comment 4: Fig.6F, lower panel: the model should show pre-tRNA, as opposed to mature tRNA, because it is the former that is fragmented.
__Response: __We apologize for the confusion. The model in Fig 7F was supposed to denote the pre-tRNA with the trailer and leader sequences intact initially, then lost with processing to mature tRNA. To make it clearer, we have now labeled the first species as “Pre-tRNA.”
Reviewer #1 (Significance (Required)):
This study is comprehensive and novel, and includes several orthogonal and complementary approaches to provide convincing evidence for the conclusions. The main discovery is significant because it presents an important advance in post-transcriptional control of gene expression. The process of tRF formation was previously thought not to affect the levels of mature tRNA. This study changes that understanding by describing for the first time the depletion of a specific mature tRNA as its precursor form is fragmented to generate tRFs. Finally, the authors identify DIS3L2 as a nuclease involved in fragmentation. This is also an important finding as the only other suspected nuclease, albeit with contradictory evidence, is angiogenin. Collectively, the findings of this study would be of interest to a broad group of scientists. I only have a few minor comments and suggestions (see above).
__Response: __We thank the reviewer for their very positive and insightful comments and feedback.
REFEREES CROSS-COMMENTING
I have the following comments on other reviewers' critiques.
Regarding the concern that the disappearance of the pre-tRNA could be a transcriptional response (reviewer 2), I think that the appearance of tRFs makes this scenario unlikely. If pre-tRNA levels decreased due to transcriptional repression, wouldn't one expect that both tRNA and the tRF levels diminish concomitantly?
Reviewer 3 raises the issue of cross hybridization in Northern blots. The authors indicate that they "could not detect the other tyrosyl tRNA (tRNA Tyr AUA) in MCF10A cells by northern blot..." (page 6). Also, they gel extracted tRFs and sequenced them (figure S6B), directly identifying the fragments. I think these findings mitigate the concern of cross hybridization and clearly identify the nature of tRFs.
Finally, I think that the codon-dependent reporter experiment (figure 5D) addresses many issues surrounding codon dependent vs indirect effects. In that experiment, the authors mutate 5 tyrosine codons of a reporter gene and demonstrate that the encoded protein is less susceptible to repression in response to oxidative stress.
__Response: __We thank the reviewer for their tremendous insights. We are in agreement regarding the three points in the cross-comments.
Reviewer #2 (Evidence, reproducibility and clarity (Required)):
This very interesting study from Sohail Tavazoie's lab describes the consequences of oxidative stress on the tRNA pool in human epithelial cell lines. As previously described, the authors observed that tRNA fragments were generated upon exposure of cells to ROS. In addition, the authors made the novel observation that specific mature tRNAs were also depleted under these conditions. In particular, the authors focused on tyrosyl tRNA-GUA, which was decreased ~50% after 24 hours of ROS exposure, an effect attributable to a decrease in the pre-tRNA pool. Depletion of tyrosyl tRNA resulted in reduced translation of specific mRNAs that are enriched in tyr codons and likely contributed to the anti-proliferative effects of ROS exposure. In addition, the authors demonstrated that the tRFs produced from tyr tRNA-GUA can interact with specific RNA binding proteins (SSB and hnRNPA1).
The major contribution of this paper is the novel finding that stress-induced tRNA fragmentation can result in a measurable reduction of specific mature tRNAs, leading to a selective reduction in translation of mRNAs that are enriched for the corresponding codons. Previously, studies of tRNA fragmentation largely focused on the functions of the tRFs themselves and it was generally believed that the mature tRNA pool was not impacted sufficiently to reduce translation. The findings reported here therefore add a new dimension to our understanding of the cellular consequences of stress-induced tRNA cleavage.
Overall, the data are of high quality, the experiments are convincing, and the conclusions are well supported. I have the following suggestions that would further strengthen the study and bolster the conclusions.
Comment 1: The authors have not formally demonstrated that the reduction in pre-tRNA in H2O2-treated cells is a consequence of pre-tRNA cleavage. It is possible that reduced transcription contributes to this effect. Pulse-chase experiments with nucleotides such as EU would provide a tractable approach to demonstrate that a labelled pool of pre-tRNA is rapidly depleted upon H2O2 treatment, which would further support their model. Since the response occurs rapidly (within 1 hour), it would be feasible to monitor the rate of pre-tRNA depletion during this time period in control vs. H2O2-treated cells.
Response: We thank the reviewer for their suggestion and agree that testing for a transcriptional effect using a pulse-chase experiment would further support these findings. We are grateful to both reviewer 1 and reviewer 2 in the cross-comments for recognizing that the tRNA repression response we see is too rapid to be a transcriptional response and that the fact that this tRNA depletion response occurs concomitantly with the tRF generation supports our model that this is a pre-tRNA fragmentation response. It would be of interest for future studies to also examine the impact of cellular stress on tRNA transcription.
Comment 2: To what extent is the growth arrest that results from H2O2 treatment attributable to tyr tRNA-GUA depletion (Fig. 3A)? Since the reduction in tRNA levels is only partial (~50%), it should be feasible to restore tRNA levels by overexpression (strategy used in Fig. 3E, S3B) and determine whether this measurably rescues growth in H2O2-treated cells.
Response: We thank the reviewer for their suggestion. Originally, we had also thought of this experiment and attempted to test this hypothesis. Upon experimentation, we ran into technical challenges that prevented us from drawing any conclusions. The problems were that we were unable to develop a cell line that stably overexpressed the Tyr tRNA-GUA and had to settle for a transient overexpression that only lasted for a couple of days (Fig S3B). For transient transfection, we used Lipofectamine 3000 (Invitrogen) that has associated cell toxicities and requires a control RNA transfection in lipofectamine. In addition, H2O2 in itself is a stress. The simultaneous occurrence of these two stresses led to a combination of cell death and cell growth for the control and experimental group. Given the high variability, we were unable to draw any conclusions on cell growth with this combination. We hope to identify a way to stably overexpress Tyr tRNA-GUA in the future to address this hypothesis.
Comment 3: Knockdown of YARS/tyr tRNA-GUA resulted in reduced expression of EPCAM, SCD, and USP3 at both the protein and mRNA levels (Fig. 4C-D, S4C). In contrast, H2O2-exposure reduced the abundance of these proteins without affecting mRNA levels (Fig. 5A-B, S5A). The authors should comment on this apparent discrepancy. Perhaps translational stalling induces No-Go decay, but it is unclear why this response would not also be triggered by ROS.
Response: We would like to clarify that out of the three genes in Fig. S5A, only EPCAM mRNA levels were significantly reduced with H2O2-exposure while no changes were observed in the mRNA levels of USP3 or SCD. It is difficult to ascertain the reason for EPCAM mRNA reduction but one hypothesis is due to timing and steady state levels. Levels of mRNAs seen with knockdown of YARS or tRNA represent steady state levels where mRNA decay and transcriptional changes can be easily seen. Following H2O2, the data is collected at 24 hours, which may be before mRNA effects can be fully appreciated. We have edited the text to clarify the uncertainty involved. We agree with the reviewer’s insightful comment and find these differences to be interesting and will consider them in future studies to better understand the interplay between translation and mRNA levels in the context of tRNA depletion.
Comment 4: In addition to the analyses of ribosome profiling in Fig. 5E-F, it might also be helpful to show a metagene analysis of ribosome occupancy centered upon UAC/UAU codons (for an example, see Figure 2 of Schuller et al., Mol Cell, 2017). This has previously been used as an effective way to visualize ribosome stalling at specific codons. Additionally, do the authors see a global correlation between tyrosine codon density and reduced translational efficiency in tRNA knockdown cells?
Response: We thank the reviewer for their important suggestion. We have expanded the analysis to look at codon usage scatterplots across all codons for shTyr and shControl replicates (Fig S5D). The 5 most changed codons are labeled with UAC, a codon for the tyrosine amino acid, being the most affected (red arrow). Consistent with our model, a tyrosine codon, when at the ribosome A-site, is most affected with depletion of the corresponding tRNA. The text has also been edited to reflect our new analysis providing further evidence that ribosomal stalling could occur upon depletion of this tRNA. The gray outline around the regression line represents the 95% confidence interval.
Fig S5D
As seen in Fig 5F, a significant overlap was noted for genes with the lowest translational efficiency and tyrosine enrichment. We did further analysis to test if a direct and linear relationship exists between tyrosine codon density and reduced translational efficiency on the global scale (i.e. does more stalling occur with more tyrosine codons on a global scale). We again see that a reduced translational efficiency is significantly correlated with tyrosine codon enrichment (above median parameters) in the tRNA knockdown ribosome profiling data. However, our analysis on a direct relationship between codon density and translational efficiency is inconclusive. This analysis is limited given the sequencing depth and number of experimental replicates available and we lack the statistical power to draw strong conclusions. To prevent overstating our claims, we have omitted any conclusions regarding this second analysis.
Comment 5: MINOR: On pg. 4, the authors state that tRF-tyrGUA is the most highly induced tRF, but Fig. S1B appears to show stronger induction of tRF-LeuTAA.
Response: The reviewer is correct in that the data from Fig S1B shows Leu-tRFs with higher induction. Our text was meant to suggest we focused on tRF-TyrGUA due to higher band intensity seen on northern blot validation. We have edited the text in the manuscript to clarify this.
Reviewer #2 (Significance (Required)):
The major advance provided by this work is the demonstration that stress-induced tRNA cleavage can reduce the abundance of the mature tRNA pool sufficiently to impact translation. Moreover, the effect on mature tRNAs is selective, resulting in the reduced translation of a specific set of mRNAs under these conditions. These findings reveal previously unknown consequences of oxidative stress on gene expression and will be of interest to scientists working on cellular stress responses and post-transcriptional regulation.
__Response: __We thank the reviewer for the kind comments and feedback.
REFEREES CROSS-COMMENTING
Regarding the concern that the disappearance of the pre-tRNA could be a transcriptional response (reviewer 2), I think that the appearance of tRFs makes this scenario unlikely. If pre-tRNA levels decreased due to transcriptional repression, wouldn't one expect that both tRNA and the tRF levels diminish concomitantly?
Here is what I was thinking: The generation of tRFs does not generally result in reduction in levels of the mature tRNAs. So you can imagine a scenario where oxidative stress causes tRF generation from the mature tyr tRNA (which does not impact its steady-state levels), as is the case for other tRNAs. At the same time, decreased transcription would reduce the pre-tRNA pool, leading to a delayed reduction in mature tRNA, as observed.
However, looking back at the data, I see that after only 5 min of H2O2 treatment, the authors observed reduced pre-tRNA and increased tRFs (Fig. 2A). This seems very fast for a transcriptional response, which would presumably require some kind of signal transduction. In addition, when you consider the amount of tRFs produced in Fig. S2C, it is hard to imagine that this would not impact the mature tRNA pool if they were derived from there. So I agree that the transcriptional scenario seems unlikely.
Nevertheless, I think that looking at pre-tRNA degradation directly with the pulse-chase strategy would strengthen their story, so I would like to give the authors this suggestion. However, I am fine with listing this as an optional experiment which would enhance the paper but should not be essential for publication.
Response: We thank the reviewer for these insightful comments. As mentioned above, five minutes is likely too rapid for a transcriptional response to be the main effect of H2O2 on Tyr-tRNA GUA. Moreover, the concomitant appearance of the tRF at this time-point makes tRNA fragmentation the most parsimonious and likely explanation rather than transcriptional repression, which would not cause a tRNA fragment to occur concurrently. Moreover, extraction and sequencing of the tRF shows it likely derives from the pre-tRNA as a 5’ leader sequence is present. We appreciate the reviewer’s suggestion and scholarly willingness to reassess their own hypothesis.
Reviewer #3 (Evidence, reproducibility and clarity (Required)):
The major findings in this manuscript are: 1.) Oxidative stress in human cells causes a decrease in tyrosine tRNA levels and accumulation of tyrosine tRNA fragments; 2.) The depletion of tyrosyl-tRNA synthetase or tyrosine tRNAs in human cells results in altered translation of certain genes and reduced cell growth and 3.) hnRNPA1 and SSB/La can bind tyrosine tRNA fragments. There is also preliminary evidence that the DIS3L2 endonuclease contributes to the appearance of tyrosine tRNA fragments upon oxidative stress. Based upon these results, the Authors conclude that tyrosine tRNA depletion is part of a conserved stress-response pathway to regulate translation in a codon-based manner.
**Major comments:**
Comment 1: There is a considerable amount of data in this paper and the experiments are performed in a generally rigorous manner. Sufficient details are provided for reproducing the findings and all results have been provided to appropriate databases (RNA-Seq and ribosome profiling).
__Response: __We thank the reviewer for the positive comments and feedback.
Comment 2: The manuscript uses a probe against the 5' half of Tyrosine tRNA for Northern blotting. However, tRNA probes can be prone to cross-hybridization, especially with some tRNA isoacceptors being similar in sequence. Thus, the blots in Figure 2 and Supplemental Figures should be probed with an oligonucleotide against the 3' half of tRNA-Tyr. This will confirm the pre- and mature tRNA-Tyr bands detected with the 5' probe. Moreover, this will determine whether 3' tRNA-Tyr fragments accumulate.
__Response: __We agree that the reviewer is correct in suggesting that the 3’ tRNA-Tyr might also accumulate. However, we disagree that any accumulation of the 3’ tRF might be relevant in our particular model for multiple reasons. As supported by reviewer 1’s cross-comments, cross-hybridization between isoacceptors (GUA vs AUA) would be unlikely as Tyr-AUA could not even be detected by the initial 5’ tRF probe. Additionally, the sequences for Tyr-GUA are different with no nucleotide alignment from Tyr-AUA. Furthermore, the extraction and sequencing of the 5’ tRF (Fig S6B) confirms the 5’ leader sequence unique to the pre-tRNA (also noted by reviewer 1). While the 3’ half of many Tyr-GUA are similar, we find selective binding of our RNA binding proteins only to the 5’ tRF. The 3’ tRF may play some role in binding to other proteins in cell regulatory pathways but such experiments would be outside the scope of this study.
__Comment 3: __The analysis of the proteomic and ribosome profiling experiments seem rather limited, or based upon what was presented in this manuscript. If additional analyses were performed, then they should be included as well, even if they yielded negative results. For example, the manuscript identifies 102 proteins that decrease after tRNA-Tyr depletion and YARS-depletion with a certain threshold of Tyr codon content. We realize the Authors were trying to find potential genes that are modulated under all three conditions. However, this does not provide information whether there is a relationship between a certain codon such as Tyr and protein abundance if only binning into two categories representing below and above a certain codon content. The Authors should plot the abundance change of each detected protein versus each codon and determine the correlation coefficient. This analysis is important for substantiating the conclusion of a codon-based system of specifically modulating transcripts enriched for certain codons. Otherwise, how could changes in tRNA-Tyr levels modulate codon-dependent gene expression if two different transcripts with the same Tyr codon content exhibit differences in translation? Moreover, this analysis should be performed with all the other codons as well.
__Response: __We have identified an error in our manuscript where the overlap identified 109 proteins and not 102 as reported previously. We apologize for this oversight. While the reviewer is correct in that identifying codon dependent changes for all 3500+ proteins detected would offer greater insight, our study was specifically focused on tyrosine as we observed this tRNA to become depleted and our experimental system modulated this specific tRNA. As for the second point on Tyr tRNA level effects on translation, we felt that the most rigorous course would be to assess causality rather than an association for this tRNA and its codon in regulating a target gene. The only way to do this is to perform mutagenesis and reporter studies. Our codon dependent reporter clearly shows a direct effect on translation in a tyrosine-codon dependent manner. As for translational regulation for two different transcripts with the same Tyr codon content, it is unclear the molecular mechanisms that could dictate these differences. The reviewer has already brought up possibilities in the next comment regarding Tyr codons in 5’ or 3’ ends or consecutive Tyr codons. These are all interesting hypotheses that others in the field have devoted entire publications to try and understand how and why codon interactions and localizations impact translation (see Gamble et al., *Cell *2016, Kunec and Osterreider, *Cell Reports *2016, Gobet et al., PNAS 2020). While these further analyses would be interesting, our current experimental data would be insufficient to properly address these questions. We have focused on a specific tRNA, its fragment, and demonstrated direct effects of the tRNA on the codon-dependent translation of a specific growth-regulating target gene and the tRNA fragment on the modulation of the activity of the RNA binding protein it binds to with respect to its regulon. We believe that these findings individually reveal causal roles for this tRNA and tRF in downstream gene regulation and collectively reveal a previously unappreciated post-transcriptional response. We hope the reviewer agrees with us regarding the already deep extent of the studies and that further such analyses beyond this tRNA are outside the scope and focus of this current study.
__Comment 4: __The Authors should provide the specific parameters used to calculate the median abundance of Tyr codons in a protein and the list of proteins containing higher than median abundance of Tyr codon content. Moreover, the complete list of 102 candidate genes should also be provided. This will allow one to determine what percentage of these Tyr-enriched proteins exhibited a decrease in levels. Moreover, is there anything special about these Tyr codon-enriched transcripts where they are affected at the level of translation but not the other Tyr-codon enriched transcripts? For example, are these transcripts enriched at the 5' or 3' ends for Tyr codons? Do these transcripts exhibit multiple consecutive Tyr codons? This deeper analysis would enrich the findings in this manuscript.
__Response: __For the proteins identified in the mass spectrometry and overlap listed in Fig 4A, Tyr codon abundance was calculated by dividing the number of Tyr amino acids present by the total number of amino acids for each protein. For genes with different isoforms possible, the principal isoform, using ENSEMBL, was used for calculations. We are also happy to provide the entire list of proteins. Additionally, please see above response to comment 3. We wish to emphasize that the goal of identification of these proteins was to identify downstream targets of this response for functional studies, which we have done. We have identified downstream genes that become modulated by this response and that regulate cell growth, consistent with the phenotype of the tRNA. We then demonstrated a direct causal tRNA-dependent codon-based response with a specific target gene using mutagenesis.
While we agree that the additional analysis the reviewer is requesting to determine what constitutes heightened translational sensitivity to this response is interesting, we believe this is a challenging question for future studies. It is possible that enrichment at 5’ or 3’ or concentration of tyrosine codons could cause increased sensitivity. Ideally, one would have information on a larger set of proteins so that such challenging questions could be better statistically bolstered. Ultimately, the requested experiments that go beyond our current work would require further analyses and experiments to allow firm conclusions to be drawn. As the other reviewers state and this reviewer agrees, we have uncovered the initial discovery regarding this tRNA fragmentation response and provided mechanistic characterization. Future studies, which are beyond the scope of the current work will undoubtedly further characterize features of this response.
__Comment 5: __The ribosome profiling results are condensed into two panels of Figure 5E and 5F. We recommend the ribosome profiling experiment be expanded into its own figure with more extensive analysis and comparison beyond just looking at tRNA-Tyr. This could reveal insight into other codons that are impacted coordinately with Tyr codons and perhaps strengthen their conclusion. As an example of a more thorough analysis of ribosome profiling and proteomics, we point the Authors to this recent paper: Lyu et al. 2020 PLoS Genetics, https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1008836
__Response: __We thank the reviewer for their suggestion. We have expanded the analysis to look at codon usage scatterplots across all codons for shTyr and shControl replicates (Fig S5D). The 5 most changed codons are labeled with UAC, a codon for the tyrosine amino acid, being the most affected (red arrow). Consistent with our model, a tyrosine codon, when at the ribosome A-site, is most affected with depletion of the corresponding tRNA. The text has also been edited to reflect our new analysis providing further evidence that ribosomal stalling might occur with depletion of a given tRNA. The gray outline around the regression line represents the 95% confidence interval.
Fig S5D
__Comment 6: __Moreover, one would expect that the mRNAs encoding USP3, EPCAM and SCD would exhibit increased ribosome occupancy. Thus, the authors should at least provide relative ribosome occupancy information on these transcripts to provide evidence that the decrease in protein levels is indeed linked to ribosome pausing or stalling.
__Response: __We would like to emphasize that resolution of ribosomal profiling data at the codon level for specific genes requires a high number of reads and replicates to draw accurate conclusions. There is an inherent level of stochasticity when mapping RPFs to specific genes and as a result, our analysis revolved around Tyr-enriched vs Tyr-low populations as this analysis was appropriate for our sequencing depth and number of replicates. To be able to conclusively make claims regarding ribosome pausing or stalling for specific genes, we would likely need further experimentation than can be currently done. However, we are currently conducting the requested bioinformatic analysis and have promising preliminary transcript-level data supporting our model.
__Comment 7: __The results with hnRNPA1 and SSB/La are extremely preliminary and simply show binding of tRNA fragments but no biological relevance. We realize that the Authors attempted to see if Tyr-tRNA fragments impacted RNA Pol III RNA but found no effect. A potential experiment would be to perform HITS-CLIP on H2O2-treated cells to see if stress-induced tRNA fragments bind to SSB/La or hnRNPA1. In this case, at least the Authors would link the oxidative stress results found in Figure 1 and 2 with La/SSB and hnRNPA1.
__Response: __We agree with the reviewer that a tRF function was not established in the manuscript. As a result, we have recently completed experiments looking at mRNA stability of the hnRNPA1 regulon in the context of overexpressing the tRF as well as using LNA to inhibit this Tyr-tRF (Fig 7E-F). Our data shows, in an hnRNPA1-dependent manner, that its regulon can be functionally regulated by Tyr-tRF. With tRF overexpression and RNAi-mediated depletion of hnRNPA1, a right shift in transcript stability is seen. Importantly, when we do the converse experiment with tRF inhibition in the same RNAi-mediated reduction of hnRNPA1, we see a left shift. These complementary experiments provide data that the Tyr-tRF has a functional role when bound to hnRNPA1 by modulating the regulon of hnRNPA1 and expand the scope of this manuscript and extend the pathway defined downstream of this tRNA fragmentation event.
Fig 7E-F
__Comment 8: __The manuscript concludes that "Tyrosyl tRNA-GUA fragments are generated in a DIS3L2-dependent manner" based upon data in Supplemental Figure S7. However, there is still a substantial amount of tyrosine tRNA fragments in both worms and human cells depleted of DIS3L2. Thus, DIS3L could play a role in the formation of Tyrosine tRNA fragments but it is too strong a claim to say that tRNA fragments are "dependent" upon DIS3L2. We suggest that the Authors soften their conclusions.
__Response: __While there are certainly tRFs still apparent with DIS3L2 depletion (Fig S7F-I), we note significant impairment of tRF induction with DIS3L2 knockdown/knockout with multiple different methods in C. elegans and human cells. This data supports our conclusion that tRF generation is dependent on DIS3L2 as this ribonuclease is necessary to elicit the full Tyr-tRF response. We do not make claims that Tyr-tRFs are solely or completely dependent on DIS3L2. There must be other RNases involved given the data highlighted by the reviewer. To this point, we have added clarifying text that DIS3L2 depletion does not completely eliminate the tRF induction.
__Comment 9: __Moreover, what is the level of DIS3L2 depletion in the worm and human cell lines? The Authors should provide the immunoblot of DIS3L2 that was described in the Materials and Methods.
__Response: __An immunoblot of DIS3L2 depletion in human cells has now been added as a supplementary figure (Fig S7I). Depletion in C. elegans was confirmed through sequencing of a mutation, as is standard in the field. The wild-type PCR product is 1nt longer (859 bp) than the mutant product (858 bp) with CTC to TAG nonsynonymous mutation preceding a single nucleotide deletion.
Wild-type disl-2: GTTGAAGCCGCAGGGC[CTC]ACTCAGACAGCTACAGG
disl-2 (syb1033): GTTGAAGCCGCAGGGC[TAG]-CTCAGACAGCTACAGG
Fig S7I
__Comment 10: __The key conclusions of "a tRNA-regulated growth suppressive oxidative stress response pathway" and an "underlying adaptive codon-based gene regulatory logic inherent to the genetic code" are overstated. This is because of the major caveat that knockdown of tyrosine-tRNA or tyrosyl-tRNA synthetase are likely to trigger numerous indirect effects. While the authors validate that three proteins are expressed at lower levels under all three conditions (H2O2, tRNA-Tyr and YARS), they might overlap in some manner but not necessarily define a coordinated response. Thus, a glaring gap in this paper is a clear, mechanistic link between H2O2-induced changes in translation versus the changes in expression when either tRNA-Tyr or YARS is depleted. Thus, it is too preliminary to conclude that tRNA depletion is part of a "pathway" and "regulatory logic" when it could all be pleiotropic effects. At the very least, the authors should discuss the possibility of indirect effects to provide a more nuanced discussion of the results obtained using two different cell systems and oxidative stress.
__Response: __We thank the reviewer for the feedback. While we agree that indirect effects may exist, we do not make any claims that our pathway is the only one required to have translation effects. The text for Fig 4A already acknowledges the pleiotropic effects of tRNA depletion. Our data shows that H2O2 stress leads to a depletion of Tyr tRNA-GUA and that depletion of this tRNA through multiple complementary methods has a codon-dependent effect on protein expression. We hope the reviewer agrees that the reduction of a specific target gene in a tyrosine codon-dependent manner (demonstrated by mutagenesis) and the binding of the tRF directly to an RBP and the modulation of the regulon of this RBP by this tRF (demonstrated by gain- and loss-of-function studies) demonstrates a direct role of this response on specific downstream target genes rather than pleiotropy. This is in keeping with the cross-comments of reviewer 1, where Fig 5D shows a direct Tyr codon link between H2O2 and downstream effects. As a result, we feel that our conclusions of a pathway (not the only pathway) are valid. However, the conclusion of a “regulatory logic” might not be interpreted in the same way by all readers and we have thus changed the text to reflect a more nuanced position.
**Minor comments:**
__Comment 11: __Tyrosyl-tRNAs refers to the aminoacylated form of tRNA. We recommend that all instances of tyrosyl-tRNA be changed to tyrosine tRNA or tRNA-Tyr which is more generic and provides no indication as to the aminoacylation status of a tRNA.
Response: We thank the reviewer for their correction. We have changed all instances of “tyrosyl” to “tyrosine” in the text.
__Comment 12: __In Figure 5C, the promoter is drawn as T7, which is a bacteriophage promoter. While the plasmid used in this manuscript (psiCHECK2) does contain a T7 promoter, mammalian gene expression is driven from the SV40 promoter. Thus, the relevant label in Figure 5C should be "SV40 promoter". Moreover, additional details should be provided on how the construct was made (such as sequence information etc.).
Response: We thank the reviewer for their correction. We have changed the promoter text in the figure. In the methods for the construct, we have included which USP3 was used and would be happy to include further information if requested.
__Comment 13: __Please provide original blots for each of the replicates in:
Figure 4C, n=4
Figure 4A, n=9
Figure 4D, n=3
Figure 5D, n=3
Response: There appears to be an unintentional mislabeling of the requested blots by the reviewer. The original blots for Fig 4C, Fig 5A, Fig 5D, and Fig 6D have been made available in a separate file for reviewers.
Reviewer #3 (Significance (Required)):
This manuscript provides evidence that specific tRNAs are depleted upon oxidative stress as part a conserved stress-response pathway in humans (and worms) to regulate translation in a codon-based manner. Unfortunately, the manuscript attempts to tie together results from different conditions and systems without providing any definitive links that suggest a "pathway" involved in the oxidative stress response. The findings in this paper provide a useful starting point but fall short of being a major advance due to the lack of a clear mechanism. However, there are intriguing results in this manuscript based upon the cell lines depleted of tRNA-Tyr or tyrosine synthetase that could interest researchers in the field of tRNA biology.
Response: We thank the reviewer for the positive comments regarding our demonstration of a conserved stress response, acknowledging the intriguing nature of our findings that will be a starting point for future studies and that our work will be of interest to researchers in the field of tRNA biology. We hope that the very positive comments of reviewer 1 and 2, the cross-comments of reviewer 1 in response to reviewer 3’s comments regarding the specificity of this response, and our inclusion for reviewer 3 of additional data on the function of the tRF in regulating the activity of the hnRNPA1 RNA binding protein defining a post-transcriptional pathway and additional corroborating requested codon-level computational analyses provide compelling support that that our findings indeed represent a major advance for the field.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
The major findings in this manuscript are: 1.) Oxidative stress in human cells causes a decrease in tyrosine tRNA levels and accumulation of tyrosine tRNA fragments; 2.) The depletion of tyrosyl-tRNA synthetase or tyrosine tRNAs in human cells results in altered translation of certain genes and reduced cell growth and 3.) hnRNPA1 and SSB/La can bind tyrosine tRNA fragments. There is also preliminary evidence that the DIS3L2 endonuclease contributes to the appearance of tyrosine tRNA fragments upon oxidative stress. Based upon these results, the Authors conclude that tyrosine tRNA depletion is part of a conserved stress-response …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #3
Evidence, reproducibility and clarity
The major findings in this manuscript are: 1.) Oxidative stress in human cells causes a decrease in tyrosine tRNA levels and accumulation of tyrosine tRNA fragments; 2.) The depletion of tyrosyl-tRNA synthetase or tyrosine tRNAs in human cells results in altered translation of certain genes and reduced cell growth and 3.) hnRNPA1 and SSB/La can bind tyrosine tRNA fragments. There is also preliminary evidence that the DIS3L2 endonuclease contributes to the appearance of tyrosine tRNA fragments upon oxidative stress. Based upon these results, the Authors conclude that tyrosine tRNA depletion is part of a conserved stress-response pathway to regulate translation in a codon-based manner.
Major comments:
•There is a considerable amount of data in this paper and the experiments are performed in a generally rigorous manner. Sufficient details are provided for reproducing the findings and all results have been provided to appropriate databases (RNA-Seq and ribosome profiling).
•The manuscript uses a probe against the 5' half of Tyrosine tRNA for Northern blotting. However, tRNA probes can be prone to cross-hybridization, especially with some tRNA isoacceptors being similar in sequence. Thus, the blots in Figure 2 and Supplemental Figures should be probed with an oligonucleotide against the 3' half of tRNA-Tyr. This will confirm the pre- and mature tRNA-Tyr bands detected with the 5' probe. Moreover, this will determine whether 3' tRNA-Tyr fragments accumulate.
•The analysis of the proteomic and ribosome profiling experiments seem rather limited, or based upon what was presented in this manuscript. If additional analyses were performed, then they should be included as well, even if they yielded negative results. For example, the manuscript identifies 102 proteins that decrease after tRNA-Tyr depletion and YARS-depletion with a certain threshold of Tyr codon content. We realize the Authors were trying to find potential genes that are modulated under all three conditions. However, this does not provide information whether there is a relationship between a certain codon such as Tyr and protein abundance if only binning into two categories representing below and above a certain codon content. The Authors should plot the abundance change of each detected protein versus each codon and determine the correlation coefficient. This analysis is important for substantiating the conclusion of a codon-based system of specifically modulating transcripts enriched for certain codons. Otherwise, how could changes in tRNA-Tyr levels modulate codon-dependent gene expression if two different transcripts with the same Tyr codon content exhibit differences in translation? Moreover, this analysis should be performed with all the other codons as well.
•The Authors should provide the specific parameters used to calculate the median abundance of Tyr codons in a protein and the list of proteins containing higher than median abundance of Tyr codon content. Moreover, the complete list of 102 candidate genes should also be provided. This will allow one to determine what percentage of these Tyr-enriched proteins exhibited a decrease in levels. Moreover, is there anything special about these Tyr codon-enriched transcripts where they are affected at the level of translation but not the other Tyr-codon enriched transcripts? For example, are these transcripts enriched at the 5' or 3' ends for Tyr codons? Do these transcripts exhibit multiple consecutive Tyr codons? This deeper analysis would enrich the findings in this manuscript.
•The ribosome profiling results are condensed into two panels of Figure 5E and 5F. We recommend the ribosome profiling experiment be expanded into its own figure with more extensive analysis and comparison beyond just looking at tRNA-Tyr. This could reveal insight into other codons that are impacted coordinately with Tyr codons and perhaps strengthen their conclusion. As an example of a more thorough analysis of ribosome profiling and proteomics, we point the Authors to this recent paper: Lyu et al. 2020 PLoS Genetics, https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1008836
•Moreover, one would expect that the mRNAs encoding USP3, EPCAM and SCD would exhibit increased ribosome occupancy. Thus, the authors should at least provide relative ribosome occupancy information on these transcripts to provide evidence that the decrease in protein levels is indeed linked to ribosome pausing or stalling.
•The results with hnRNPA1 and SSB/La are extremely preliminary and simply show binding of tRNA fragments but no biological relevance. We realize that the Authors attempted to see if Tyr-tRNA fragments impacted RNA Pol III RNA but found no effect. A potential experiment would be to perform HITS-CLIP on H2O2-treated cells to see if stress-induced tRNA fragments bind to SSB/La or hnRNPA1. In this case, at least the Authors would link the oxidative stress results found in Figure 1 and 2 with La/SSB and hnRNPA1.
•The manuscript concludes that "Tyrosyl tRNA-GUA fragments are generated in a DIS3L2-dependent manner" based upon data in Supplemental Figure S7. However, there is still a substantial amount of tyrosine tRNA fragments in both worms and human cells depleted of DIS3L2. Thus, DIS3L could play a role in the formation of Tyrosine tRNA fragments but it is too strong a claim to say that tRNA fragments are "dependent" upon DIS3L2. We suggest that the Authors soften their conclusions.
•Moreover, what is the level of DIS3L2 depletion in the worm and human cell lines? The Authors should provide the immunoblot of DIS3L2 that was described in the Materials and Methods.
•The key conclusions of "a tRNA-regulated growth suppressive oxidative stress response pathway" and an "underlying adaptive codon-based gene regulatory logic inherent to the genetic code" are overstated. This is because of the major caveat that knockdown of tyrosine-tRNA or tyrosyl-tRNA synthetase are likely to trigger numerous indirect effects. While the authors validate that three proteins are expressed at lower levels under all three conditions (H2O2, tRNA-Tyr and YARS), they might overlap in some manner but not necessarily define a coordinated response. Thus, a glaring gap in this paper is a clear, mechanistic link between H2O2-induced changes in translation versus the changes in expression when either tRNA-Tyr or YARS is depleted. Thus, it is too preliminary to conclude that tRNA depletion is part of a "pathway" and "regulatory logic" when it could all be pleiotropic effects. At the very least, the authors should discuss the possibility of indirect effects to provide a more nuanced discussion of the results obtained using two different cell systems and oxidative stress.
Minor comments:
•Tyrosyl-tRNAs refers to the aminoacylated form of tRNA. We recommend that all instances of tyrosyl-tRNA be changed to tyrosine tRNA or tRNA-Tyr which is more generic and provides no indication as to the aminoacylation status of a tRNA.
•In Figure 5C, the promoter is drawn as T7, which is a bacteriophage promoter. While the plasmid used in this manuscript (psiCHECK2) does contain a T7 promoter, mammalian gene expression is driven from the SV40 promoter. Thus, the relevant label in Figure 5C should be "SV40 promoter". Moreover, additional details should be provided on how the construct was made (such as sequence information etc.).
•Please provide original blots for each of the replicates in:
Figure 4C, n=4
Figure 4A, n=9
Figure 4D, n=3
Figure 5D, n=3
Significance
This manuscript provides evidence that specific tRNAs are depleted upon oxidative stress as part a conserved stress-response pathway in humans (and worms) to regulate translation in a codon-based manner. Unfortunately, the manuscript attempts to tie together results from different conditions and systems without providing any definitive links that suggest a "pathway" involved in the oxidative stress response. The findings in this paper provide a useful starting point but fall short of being a major advance due to the lack of a clear mechanism. However, there are intriguing results in this manuscript based upon the cell lines depleted of tRNA-Tyr or tyrosine synthetase that could interest researchers in the field of tRNA biology.
This review is written from the perspective of a researcher with expertise in RNA processing, RNA biology and translation regulation.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
This very interesting study from Sohail Tavazoie's lab describes the consequences of oxidative stress on the tRNA pool in human epithelial cell lines. As previously described, the authors observed that tRNA fragments were generated upon exposure of cells to ROS. In addition, the authors made the novel observation that specific mature tRNAs were also depleted under these conditions. In particular, the authors focused on tyrosyl tRNA-GUA, which was decreased ~50% after 24 hours of ROS exposure, an effect attributable to a decrease in the pre-tRNA pool. Depletion of tyrosyl tRNA resulted in reduced translation of specific mRNAs that …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
This very interesting study from Sohail Tavazoie's lab describes the consequences of oxidative stress on the tRNA pool in human epithelial cell lines. As previously described, the authors observed that tRNA fragments were generated upon exposure of cells to ROS. In addition, the authors made the novel observation that specific mature tRNAs were also depleted under these conditions. In particular, the authors focused on tyrosyl tRNA-GUA, which was decreased ~50% after 24 hours of ROS exposure, an effect attributable to a decrease in the pre-tRNA pool. Depletion of tyrosyl tRNA resulted in reduced translation of specific mRNAs that are enriched in tyr codons and likely contributed to the anti-proliferative effects of ROS exposure. In addition, the authors demonstrated that the tRFs produced from tyr tRNA-GUA can interact with specific RNA binding proteins (SSB and hnRNPA1).
The major contribution of this paper is the novel finding that stress-induced tRNA fragmentation can result in a measurable reduction of specific mature tRNAs, leading to a selective reduction in translation of mRNAs that are enriched for the corresponding codons. Previously, studies of tRNA fragmentation largely focused on the functions of the tRFs themselves and it was generally believed that the mature tRNA pool was not impacted sufficiently to reduce translation. The findings reported here therefore add a new dimension to our understanding of the cellular consequences of stress-induced tRNA cleavage.
Overall, the data are of high quality, the experiments are convincing, and the conclusions are well supported. I have the following suggestions that would further strengthen the study and bolster the conclusions.
1.The authors have not formally demonstrated that the reduction in pre-tRNA in H2O2-treated cells is a consequence of pre-tRNA cleavage. It is possible that reduced transcription contributes to this effect. Pulse-chase experiments with nucleotides such as EU would provide a tractable approach to demonstrate that a labelled pool of pre-tRNA is rapidly depleted upon H2O2 treatment, which would further support their model. Since the response occurs rapidly (within 1 hour), it would be feasible to monitor the rate of pre-tRNA depletion during this time period in control vs. H2O2-treated cells.
2.To what extent is the growth arrest that results from H2O2 treatment attributable to tyr tRNA-GUA depletion (Fig. 3A)? Since the reduction in tRNA levels is only partial (~50%), it should be feasible to restore tRNA levels by overexpression (strategy used in Fig. 3E, S3B) and determine whether this measurably rescues growth in H2O2-treated cells.
3.Knockdown of YARS/tyr tRNA-GUA resulted in reduced expression of EPCAM, SCD, and USP3 at both the protein and mRNA levels (Fig. 4C-D, S4C). In contrast, H2O2-exposure reduced the abundance of these proteins without affecting mRNA levels (Fig. 5A-B, S5A). The authors should comment on this apparent discrepancy. Perhaps translational stalling induces No-Go decay, but it is unclear why this response would not also be triggered by ROS.
4.In addition to the analyses of ribosome profiling in Fig. 5E-F, it might also be helpful to show a metagene analysis of ribosome occupancy centered upon UAC/UAU codons (for an example, see Figure 2 of Schuller et al., Mol Cell, 2017). This has previously been used as an effective way to visualize ribosome stalling at specific codons. Additionally, do the authors see a global correlation between tyrosine codon density and reduced translational efficiency in tRNA knockdown cells?
5.MINOR: On pg. 4, the authors state that tRF-tyrGUA is the most highly induced tRF, but Fig. S1B appears to show stronger induction of tRF-LeuTAA.
Significance
The major advance provided by this work is the demonstration that stress-induced tRNA cleavage can reduce the abundance of the mature tRNA pool sufficiently to impact translation. Moreover, the effect on mature tRNAs is selective, resulting in the reduced translation of a specific set of mRNAs under these conditions. These findings reveal previously unknown consequences of oxidative stress on gene expression and will be of interest to scientists working on cellular stress responses and post-transcriptional regulation.
REFEREES CROSS-COMMENTING
Regarding the concern that the disappearance of the pre-tRNA could be a transcriptional response (reviewer 2), I think that the appearance of tRFs makes this scenario unlikely. If pre-tRNA levels decreased due to transcriptional repression, wouldn't one expect that both tRNA and the tRF levels diminish concomitantly?
Here is what I was thinking: The generation of tRFs does not generally result in reduction in levels of the mature tRNAs. So you can imagine a scenario where oxidative stress causes tRF generation from the mature tyr tRNA (which does not impact its steady-state levels), as is the case for other tRNAs. At the same time, decreased transcription would reduce the pre-tRNA pool, leading to a delayed reduction in mature tRNA, as observed.
However, looking back at the data, I see that after only 5 min of H2O2 treatment, the authors observed reduced pre-tRNA and increased tRFs (Fig. 2A). This seems very fast for a transcriptional response, which would presumably require some kind of signal transduction. In addition, when you consider the amount of tRFs produced in Fig. S2C, it is hard to imagine that this would not impact the mature tRNA pool if they were derived from there. So I agree that the transcriptional scenario seems unlikely.
Nevertheless, I think that looking at pre-tRNA degradation directly with the pulse-chase strategy would strengthen their story, so I would like to give the authors this suggestion. However, I am fine with listing this as an optional experiment which would enhance the paper but should not be essential for publication.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The manuscript by Huh et al. reports that oxidative stress causes fragmentation of a specific tyrosine pre-tRNA, leading to two parallel outcomes. First, the fragmentation depletes the mature tRNA, causing translational repression of genes that are disproportionally rich in tyrosine codon. These genes are enriched for those involved in electron transport chain, cell cycle and growth. Second, the fragmentation generates tRNA fragments (tRFs) that bind to two known RNA binding proteins. Finally, the authors identify a nuclease that is needed for efficient formation of tyrosine tRFs.
The authors should include a short diagram …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The manuscript by Huh et al. reports that oxidative stress causes fragmentation of a specific tyrosine pre-tRNA, leading to two parallel outcomes. First, the fragmentation depletes the mature tRNA, causing translational repression of genes that are disproportionally rich in tyrosine codon. These genes are enriched for those involved in electron transport chain, cell cycle and growth. Second, the fragmentation generates tRNA fragments (tRFs) that bind to two known RNA binding proteins. Finally, the authors identify a nuclease that is needed for efficient formation of tyrosine tRFs.
The authors should include a short diagram indicating the various known steps of pre-tRNA fragmentation (perhaps as a supplement) for general readers.
I find the enrichment for mitochondrial electron transport chain (ETC) curious. The ETC includes several oxidoreductases, which may be rich in tyrosine as it is a common amino acid used in electron transfer. The depletion of the tyrosine tRNA from among many tRNAs under oxidative stress may not be incidental but related to an attempt by the cell to decrease oxygen consumption to avoid further oxidative damage. The authors could further mine their data to corroborate this hypothesis. For example, are the ETC genes among the targets of the RNA binding proteins targeted by tyrosine tRFs? This could potentially connect the effects of mature tRNA depletion and tRFs.
In figure 4A, the authors should provide the tyrosine codon content of the overlap genes and show how much it differs from a randomly selected sample.
Fig.6F, lower panel: the model should show pre-tRNA, as opposed to mature tRNA, because it is the former that is fragmented.
Significance
This study is comprehensive and novel, and includes several orthogonal and complementary approaches to provide convincing evidence for the conclusions. The main discovery is significant because it presents an important advance in post-transcriptional control of gene expression. The process of tRF formation was previously thought not to affect the levels of mature tRNA. This study changes that understanding by describing for the first time the depletion of a specific mature tRNA as its precursor form is fragmented to generate tRFs. Finally, the authors identify DIS3L2 as a nuclease involved in fragmentation. This is also an important finding as the only other suspected nuclease, albeit with contradictory evidence, is angiogenin. Collectively, the findings of this study would be of interest to a broad group of scientists. I only have a few minor comments and suggestions (see above).
REFEREES CROSS-COMMENTING
I have the following comments on other reviewers' critiques.
Regarding the concern that the disappearance of the pre-tRNA could be a transcriptional response (reviewer 2), I think that the appearance of tRFs makes this scenario unlikely. If pre-tRNA levels decreased due to transcriptional repression, wouldn't one expect that both tRNA and the tRF levels diminish concomitantly?
Reviewer 3 raises the issue of cross hybridization in Northern blots. The authors indicate that they "could not detect the other tyrosyl tRNA (tRNA Tyr AUA) in MCF10A cells by northern blot..." (page 6). Also, they gel extracted tRFs and sequenced them (figure S6B), directly identifying the fragments. I think these findings mitigate the concern of cross hybridization and clearly identify the nature of tRFs.
Finally, I think that the codon-dependent reporter experiment (figure 5D) addresses many issues surrounding codon dependent vs indirect effects. In that experiment, the authors mutate 5 tyrosine codons of a reporter gene and demonstrate that the encoded protein is less susceptible to repression in response to oxidative stress.
-
-