mRNA Imprinting: transcription apparatus can remotely control cytoplasmic post-transcriptional mechanisms by dozens of proteins
Curation statements for this article:-
Curated by eLife
eLife Assessment
In this valuable study, the authors develop new approaches to investigate mRNA imprinting, a phenomenon in which RNA-protein complexes form in the nucleus to influence the fate of transcripts in the cytoplasm. They propose that the Pol II subunit Rpb4 serves as a key node in this pathway, recruiting proteins involved in cytoplasmic processes. Notably, some of the candidates identified in this study were previously thought to function exclusively in the cytoplasm. However, the evidence remains incomplete, as key controls are lacking and alternative explanations have not been fully addressed; additional validation would help strengthen the authors' conclusions.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
Proper regulation of gene expression requires the harmonious coordination of transcription with other stages of the mRNA lifecycle. We have previously demonstrated that proteins can bind to nascent transcripts co-transcriptionally – a phenomenon named “mRNA imprinting” because the bound proteins modulate subsequent stages of the mRNA lifecycle. Using a novel high-throughput approach, “PRofiling OF Imprinted Transcripts” (“PROFIT”), to identify proteins involved in mRNA imprinting, we uncovered several dozen candidates. The RNA polymerase II (Pol II) subunit Rpb4, which can itself imprint mRNAs, mediates the imprinting of a large subset of these proteins. Imprinting repertoire and profile is responsive to environmental changes and include HSP70 variants. Interestingly, PROFIT identified proteins previously thought to function mainly or exclusively in the cytoplasm, including a translation factors, mRNA decay factor, protein chaperones, substrate-delivering factors of the proteasome, targeting factors that deliver mRNAs to mitochondria and more. Using proximity labeling, we validated several hits, including two translation initiation factors: eIF4G and the eIF3 component Rpg1. Importantly, we found that the same Rpg1 molecule, which had been transiently localized near Pol II, co-sediments with polyribosomes - similar to the bulk Rpg1. Our results suggest that the transcription machinery can regulate translation by recruiting specific translation factors, which later participate in protein synthesis. mRNA imprinting appears to be a widespread phenomenon, and we speculate that it may not be limited to the transcription stage alone. Interestingly, the PROFIT experiments identified proteins with known cytoplasmic function, including a translation factor (Rpg1), mRNA decay factor (Xrn1), protein chaperones (Ssa1/2), substrate-delivering factors of the proteasome, targeting factors that deliver mRNAs to mitochondria, and actin-binding factors.
Article activity feed
-
eLife Assessment
In this valuable study, the authors develop new approaches to investigate mRNA imprinting, a phenomenon in which RNA-protein complexes form in the nucleus to influence the fate of transcripts in the cytoplasm. They propose that the Pol II subunit Rpb4 serves as a key node in this pathway, recruiting proteins involved in cytoplasmic processes. Notably, some of the candidates identified in this study were previously thought to function exclusively in the cytoplasm. However, the evidence remains incomplete, as key controls are lacking and alternative explanations have not been fully addressed; additional validation would help strengthen the authors' conclusions.
-
Reviewer #1 (Public review):
Summary:
To understand the process of mRNA imprinting, the authors develop a series of unbiased methods to identify and follow proteins that associate with transcripts co-transcriptionally. The methods rely on RNA polymerase II pull-downs or proximity biotinylation to do so, and from these experiments, the authors identify some interesting candidate proteins, including Rpg1 / eIF3a, Ssa1/2, and Spt6. The authors characterize some of these proteins in follow-up experiments and show that Spt6 recruitment depends on Rpb4.
Strengths:
(1) The methods described in this study will be useful for the community beyond their immediate application.
(2) The topic of mRNA imprinting remains an open area in the field, and this paper provides hypothesis-generating datasets that may be of use.
(3) If correct, the idea that …
Reviewer #1 (Public review):
Summary:
To understand the process of mRNA imprinting, the authors develop a series of unbiased methods to identify and follow proteins that associate with transcripts co-transcriptionally. The methods rely on RNA polymerase II pull-downs or proximity biotinylation to do so, and from these experiments, the authors identify some interesting candidate proteins, including Rpg1 / eIF3a, Ssa1/2, and Spt6. The authors characterize some of these proteins in follow-up experiments and show that Spt6 recruitment depends on Rpb4.
Strengths:
(1) The methods described in this study will be useful for the community beyond their immediate application.
(2) The topic of mRNA imprinting remains an open area in the field, and this paper provides hypothesis-generating datasets that may be of use.
(3) If correct, the idea that eIF3a binds co-transcriptionally would be of interest to the transcription and translation fields.
(4) The data showing the importance of Rpb4 for Spt6 binding are some of the strongest.
Weaknesses:
(1) Two main methods (PROFIT and BioPROFIT) are introduced in this study, both of which make use of a combination of tags, especially on RNA polymerase II subunits, to identify and track proteins that are potentially recruited co-transcriptionally. However, a more thorough characterization is needed to gain a sense of the false negatives and false positives. For instance, there are no direct experiments testing the requirement for transcription for the hits. This is a key experiment.
(2) Alternatives are also not robustly considered. For example, what is the evidence that the proteins remain bound to an RNA through its life cycle, as opposed to rebinding in the cytoplasm? For proteins with known cytoplasmic functions, like Rpg1/eIF3a, this conclusion needs more supporting evidence. This caveat is especially important to consider given the typical or known off-rates of many of these proteins.
(3) Showing direct evidence that biotinylated "target" proteins (like eIF3a) accumulate in the nucleus during short labeling or if nuclear export is blocked is an important control, as is an experiment inhibiting transcription and demonstrating that the signal decreases.
-
Reviewer #2 (Public review):
Summary:
The authors have provided valuable and solid evidence for the hypothesis, of which Choder is an early advocate, that transcription facilitates the assembly of an mRNA-protein complex that can affect the expression of mRNA (e.g., translation or degradation) in the cytoplasm.
Strengths:
In this work the authors have used two orthogonal approaches: an IP of a Flag labeled Pol II and RNAse digestion to release nascent chain associated proteins followed by mass spectrometry to identify cotranscriptional-associated proteins and then verifying this association with the transcriptional apparatus by proximity labeling technology using biotinylation of a specific sequence (Avi-tag) by the bacterial enzyme, BirA fused to a subunit of Pol II. Many of the proteins identified are thought to be exclusively …
Reviewer #2 (Public review):
Summary:
The authors have provided valuable and solid evidence for the hypothesis, of which Choder is an early advocate, that transcription facilitates the assembly of an mRNA-protein complex that can affect the expression of mRNA (e.g., translation or degradation) in the cytoplasm.
Strengths:
In this work the authors have used two orthogonal approaches: an IP of a Flag labeled Pol II and RNAse digestion to release nascent chain associated proteins followed by mass spectrometry to identify cotranscriptional-associated proteins and then verifying this association with the transcriptional apparatus by proximity labeling technology using biotinylation of a specific sequence (Avi-tag) by the bacterial enzyme, BirA fused to a subunit of Pol II. Many of the proteins identified are thought to be exclusively cytoplasmic, for instance, those important for translation, such as the components of initiation factor EF3. The work represents a significant advance in support of the model where specific mRNAs can assemble proteins needed for their function in the cytoplasm during their transcription.
They also discover that a mutant Pol II subunit, Rbp4, which does not bind certain Avi-tagged proteins, does not facilitate their biotinylation. These results lend credible support to the hypothesis.
Weaknesses:
While the proximity labeling provides strong evidence that is consistent with the hypothesis, a proof is still lacking because it is inferred that the enzymatic labeling occurs at the site of transcription (a reasonable assumption). More definitive evidence could be provided by imaging the presence of the cytoplasmic proteins at the transcription site, although this may not be within the expertise of the investigator, so it would require a collaboration.
While not necessarily a significant weakness, it is worth considering that a remote possibility is that the cytoplasmic proteins discovered in this way were not tagged with biotin in the nucleus, but rather in the cytoplasm, where the Pol II-complex, either Flag or BirA tagged, may come in contact with the substrate before it is imported to the nucleus. The authors presumably rule out that the tagging could occur during translation of the Avi-tag on polysomes by inhibiting translation and showing that the tagging of the target protein is not inhibited (the data here is not totally convincing). Whether the Pol II-(BirA or Flag) could react with Avi-tagged proteins, even while briefly in the cytoplasm before nuclear import, is not completely resolved by these experiments since the Avi-tagged proteins could reside in the cytoplasm, not associated with polysomes, but complexed with Pol II subunits. The mutant Rpb does not rule out this possibility since it would not bind its substrate in the cytoplasm. In order to get into the nucleus in the first place, the cytoplasmic proteins would need to be transported there by a complex, possibly involving Pol II subunits, Rpbs. Perhaps the authors could address this possibility in the text.
One confusing issue in the protocol is the efficacy of the biotin-depleted media in which the cells are grown. Biotin is an essential cofactor for many reactions, so there are still endogenous biotin and biotin ligase needed that may add a background level of promiscuous biotinylation of some cytoplasmic proteins, for instance, those containing a universal biotin binding site.
-
Reviewer #3 (Public review):
Summary:
Various groups over the last several decades have provided many examples of proteins associating with nascent mRNA co-transcriptionally to influence gene expression at subsequent stages, including in the cytoplasm. In this and previously published works, the Choder group has described these events as "mRNA imprinting", which we know as a field that reflects the differential association of proteins with mRNAs in a gene-specific or environmentally induced fashion to regulate gene expression.
In this study, the authors use a proteomics-based approach termed PROFIT to identify factors associated with RNA Pol II in an RNA-dependent manner. The identified interactors have the potential to be part of mRNA-protein complexes (mRNPs) being formed co-transcriptionally with an "mRNA imprinting" function. PROFIT …
Reviewer #3 (Public review):
Summary:
Various groups over the last several decades have provided many examples of proteins associating with nascent mRNA co-transcriptionally to influence gene expression at subsequent stages, including in the cytoplasm. In this and previously published works, the Choder group has described these events as "mRNA imprinting", which we know as a field that reflects the differential association of proteins with mRNAs in a gene-specific or environmentally induced fashion to regulate gene expression.
In this study, the authors use a proteomics-based approach termed PROFIT to identify factors associated with RNA Pol II in an RNA-dependent manner. The identified interactors have the potential to be part of mRNA-protein complexes (mRNPs) being formed co-transcriptionally with an "mRNA imprinting" function. PROFIT employs a pulldown of RNA Pol II via a tagged Rpb3 subunit, followed by RNase I-mediated elution to isolate proteins associated in an RNA-dependent manner. Proteomics analyses identified known mRNA-associated proteins that have previously been reported as imprinting factors, as well as other proteins involved in gene expression, including factors functioning in the cytoplasm. The authors suggest, based on the RNA-dependence and assumed formation of these interactions with RNA Pol II co-transcriptionally, that these novel hits could be mRNA imprinting factors. Although for most of these factors, it has not been determined whether they associate with RNA-Pol II in the context of transcription with nascent transcripts to contribute to the downstream regulations of these transcripts.
Strengths:
PROFIT successfully identified nuclear factors known to engage mRNA co-transcriptionally. This suggests that the method has the potential to detect imprinting factors. By employing a proximity-labeling technique, termed BioPROFIT, further evidence is provided for some of the novel interactors being in proximity to RNA Pol II. The authors further demonstrate that one of the factors, the eIF3 component Rpg1, exists in two fractions, with a soluble fraction that matures into a ribosome fraction, which is suggestive of Rpg1 traveling along the gene expression pathway with an mRNP to be engaged in translation. In addition, the authors showed that PROFIT detects changes in RNA Pol II associated factors in response to heat shock, consistent with gene expression reprogramming during stress. As such, these methods and proteomics data provide a starting point for a more detailed characterization of mRNP compositions formed in the nucleus and their impact on gene expression at later stages.
Weaknesses:
The authors interpret the interaction data from PROFIT and BioPROFIT under the assumption that this reflects interactions happening co-transcriptionally. There is no discussion of other ways these data may result, or more importantly, controls to prove these assumptions are true. Overall, these assays lack important controls and experimental validations by independent methods to demonstrate that the identified interactions occur co-transcriptionally within the nucleus and do not represent interactions occurring in the cytoplasm or artifacts related to experimental design. For example, the authors focus on Rpg1 as a potential imprinting factor, which would require this protein to shuttle and be localized at transcribing genes. Yet no evidence is presented that Rpg1 enters the nucleus or can be found in association with a transcribed gene, which leaves open the possibility that this interaction is occurring in the cytoplasm or forming post-lysis.
To the possibility of in vitro interactions, in the PROFIT assay, yeast collected from a 3L culture is cryo-ground and resuspended in 7 mL of lysis buffer. This ratio of cell material to buffer will create a highly concentrated cell lysate that is subsequently used over ~6.5 hours, which is the time for centrifugation, DNase I digestion, and immunoprecipitation. These conditions have a very high probability of promoting new interactions between RNA, RNA Poll II, other proteins, and/or RNA Pol II-associated nascent RNA complexes in vitro. Notably, the PROFIT assay detects many highly expressed proteins but does not identify many of the factors known to be loaded into nuclear mRNPs (e.g., Yra1, THO complex, Sub2, or Nab2). The BioPROFIT assay is used to try to address this issue, but biotinylation may occur post-lysis because the desalting process to remove biotin is performed just before the immunoprecipitation, providing ~2 hours for the reaction to happen in vitro. In addition, even if the biotinylation occurs in cells, nothing about this assay indicates this is occurring in the context of transcribing RNA Pol II or nascent transcripts. To address this major issue, the authors should add a mixing control to show that the detected interactions between RNA Pol II and the identified factors are produced in cells, not in the cell lysate. Specifically, mixing cell grindates from two independent yeast strains (e.g., RPB3-FLAG strain mixed with a TIF4631-HA strain) with the lysate used in the PROFIT assay with western blotting. In this case, if the interaction is detected, the interaction is produced in the cell lysate. To verify PROFIT hits associated with transcribing RNA Pol II and nascent transcripts, BIOPROFIT should be performed in cells treated with a transcription inhibitor (e.g., thiolutin) or mutants blocking transcription by Pol II. These types of verifications should be performed for the multiple novel hits reported in the manuscript.
Another in vitro issue must also be addressed. In the PROFIT assay, elution of RNA-associated factors from the immunoprecipitated material is performed by RNase I digestion, but the reaction time is very long (3 hours) at room temperature. During such a long incubation time and at higher temperature (i.e., above 4 Celsius), it is possible that non-RNA-mediated interactors dissociate from the beads and/or protein binding partners. This possibility is made more problematic by the fact that the authors define interactors using fold change over an Rpb3 no tag sample, where the sample does not contain isolated RNA Pol II complexes and their associated protein-binding partners. As such, even a small amount of non-RNA-mediated RNA Poll II interactors that elute would appear significantly enriched. For this point, a comparison of +/- RNase I elution in the Rpb3-FLAG pulldown sample should be performed using PROFIT.
Other points to address:
(1) The cartoon in Figure 1A should be corrected to present the PROFIT experiment as described in the text. Specifically, in the cartoon, UV is shown to be applied to cells, but this is done with cell grindate.
(2) The cartoon in Figure 2A should be corrected. In the cartoon, it shows the biotin ligase biotinylating proximal proteins during DNase digestion as well as on the Sepharose beads, but in theory, the majority of the biotinylation reaction occurs in cells. In addition, the cartoon depicts biotinylation of proximal proteins, but the system described uses wild-type BirA to specifically biotinylate an Avi-tag. To perform non-specific labeling of proximal proteins, BirA* would need to be used. Finally, the cartoon indicates mass spectrometry analysis of labeled proteins, but this is not done in the manuscript.
(3) In the text, the sentence "However, no bio-Spt6-Avi was released from the complexes containing Pol II mutants (Fig. 5C)" appears to have two errors. "Pol II mutants" should likely be "rpb4 mutant" and "Fig. 5C" is probably "Fig. 6C".
(4) In the Figure 6 legend, the sentence "The bulk Spt6 was detected by anti-HIS Abs that bound to (HIS)x6, which was placed upstream of the FLAG" suggests that "FLAG" should be "Avi-tag." Please correct it if necessary and accurately describe it in the strain list.
(5) On page 18, Npl3 is listed and discussed, but never mentioned anywhere prior in the paper. For example, the paragraph states "...our observation that it binds nascent RNA in an Rpb4-dependent manner...", but Npl3 is not listed in the supplemental Table 4, which lists PROFIT hits affected by rpb4∆. If Npl3 is to be discussed, the associated data needs to be properly presented.
-
