Fission yeast Caprin protein is required for efficient heterochromatin establishment

  1. Haidao Zhang
  2. Ekaterina Kapitonova
  3. Adriana Orrego
  4. Christos Spanos
  5. Joanna Strachan
  6. Elizabeth H. Bayne

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Heterochromatin is a key feature of eukaryotic genomes that serves important regulatory and structural roles in regions such as centromeres. In fission yeast, maintenance of existing heterochromatic domains relies on positive feedback loops involving histone methylation and non-coding RNAs. However, requirements for de novo establishment of heterochromatin are less well understood. Here, through a cross-based assay we have identified a novel factor influencing the efficiency of heterochromatin establishment. We determine that the previously uncharacterised protein is an ortholog of human Caprin1, an RNA-binding protein linked to stress granule formation. We confirm that the fission yeast ortholog, here named Cpn1, also associates with stress granules, and we uncover evidence of interplay between heterochromatin integrity and ribonucleoprotein (RNP) granule formation, with heterochromatin mutants showing reduced granule formation in the presence of stress, but increased granule formation in the absence of stress. We link this to regulation of non-coding heterochromatic transcripts, since in heterochromatin-deficient cells, absence of Cpn1 leads to hyperaccumulation of centromeric RNAs at centromeres. Together, our findings unveil a novel link between RNP homeostasis and heterochromatin assembly, and implicate Cpn1 and associated factors in facilitating efficient heterochromatin establishment by enabling removal of excess transcripts that would otherwise impair assembly processes.

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    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

    This study identified Cpn1, a fission yeast ortholog of human Caprin1, to be involved in heterochromatin re-establishment in S. pompe by potentially regulating heterochromatic transcript stability and localization. Moreover, Cpn1 was shown to be important for stress granule formation. Although the role of Cpn1 for heterochromatin establishment and granule formation is strong, the point that there is a crosstalk between the two is weaker and could be explained through multiple independent effects. Overall, the manuscript provides interesting new observations with regards to Cpn1 function that are adequate for publication, however, a few additional validations need to be performed to strengthen the crosstalk conclusion, which could be a new significant connection.

    The following comments could be addressed before publication to improve the manuscript.

    Specific Comments

    1. The point that there is a "Crosstalk between heterochromatin integrity and cytoplasmic RNP granule formation" could be strengthened before publication, or the tone of the manuscript revised to reflect the issues below.

    A) There is no data in this section showing a direct crosstalk between heterochromatin integrity and granules. The results show that certain mutants alter heterochromatin integrity the formation of PABP-containing granules. However, those proteins could have different functions in the nucleus, cytoplasm and upon stress. Thus, granule formation could be independent of heterochromatin perturbation. E.g. Loss of Ago1 could impact cytoplasmic RNA abundance/ stability and this therefore influence granule assembly. Similarly, Cpn1/Caprin could be a multifunctional protein in S. pombe and affect various aspects of RNA metabolism. The possibility, that these proteins have multiple different functions in the cell and heterochromatin integrity and cytoplasmic RNP granule formation could be due to different functions of those proteins and not due to a crosstalk should be discussed as well.

    B) Figure 5 title says that "disruption of heterochromatin alters the formation of PABP-containing RNP granules". There is no data in this figure that shows that "disruption of heterochromatin" directly causes granules. Its rather that heterochromatin mutants show altered PABP-containing RNP granules. While this is fine, it should be pointed out that this is a correlation, with a direct connection being inferred.

    C) The strongest connection of heterochromatin integrity to RNP granules is the accumulation of heterochromatic transcripts in those granules. Therefore, the manuscript could be strengthened by rearranging the sections/figures.

    In addition, Fig. 5A shows PABP containing granules in unstressed conditions for rik1, clr4, ago1 loss. This suggests that those granules contain lots of polyadenylated RNA. Evidence is needed that the mutations studied are not affecting global cytoplasmic translation or mRNA decay (e.g. by puromycin staining and smRNA-FISH staining or qPCR (e.g. GAPDH)). Moreover, it needs to be shown that endogenous expression of Cpn1 is unchanged. If those perturbations affect Cpn1 (or Nxt3) levels, the granule phenotype could be solely due to changes in stress granule promoting proteins.

    1. The work would be strengthened by adding some additional experiments. Specifically:

    A) "Previous studies by ourselves and others have provided evidence that accumulation of transcripts on chromatin can impair heterochromatin assembly, possibly through increased formation of RNA-DNA hybrids". Increased RNA-DNA hybrids/ R-loop structures were shown to lead to genomic instability, which could lead to micronuclei formation and nuclei leakage in the cytoplasm during stress. It would be nice to provide close-up view images of those stress induced cytoplasmic granules, that contain cenRNA. Do they contain weaker DAPI signal in those granules, which would be indicative of micronuclei? Moreover, providing an additional stain using either an R-loop antibody, cGAS, or a nuclei membrane marker such as LAMIN B1 could be used to rule out micronuclei/ membranous assemblies.

    B) RNA FISH and quantification for PABP foci in unstressed and stressed clr4Δ dhp1-2 cells is key, which should show the highest changes in foci and RNA accumulation in foci. Are those cells showing an even stronger effect on heterochromatin establishment?

    C) Compared to clr4Δ cpn1Δ, do clr4Δ dhp1-2 cpn1Δ (or cpn1Δ and dhp1-2 cpn1Δ) form more stress induced granules? And do those granules contain more heterochromatic transcripts?

    D) Is heterochromatic transcripts localization in -glucose induced granules also seen with heat shock?

    1. Additional tests and quantification is needed to support that conclusion: "...whereas heterochromatin mutants show increased Pabp granule formation in absence of stress, they show a significant reduction in the average number of Pabp foci formed in the presence of stress,...". Different proteins could regulate/loss of proteins impact granule assembly in different manners, for example RNA localization, assembly, disassembly, foci number, foci size, % cells with foci etc. It looks like rik1Δ shows larger foci. Therefore, upon stress, there could be indeed fewer foci because they are larger. A quantification of foci area and total foci area per cell should support or reject that conclusion. Moreover, is the same trend also observed with starvation stress, or only heat shock?
    2. The field generally believes G3BP1 is not an endonuclease, and therefore the authors might want to edit the section: " ...these could include, for example, Cpn1 binding partner Nxt3, the human ortholog of which, G3BP1, has been shown to function as an endoribonuclease for degradation of selected RNAs (63)." Although, it was shown that G3BP1 has endoribonuclease activity in that reference, this was never reproduced and is generally now accepted in the field that G3BP1 does not function as a endoribonuclease to regulate RNA homeostasis.

    Minor comment:

    1. Fig. 2 is a bit confusing. Maybe it could help if the controls - to rule out heterochromatin maintenance (B, C,D) - could be better grouped together or put into supplementary.
    2. Fig 4 A, figures for Cpn1R6A-GFP upon glucose starvation is missing.
    3. Fig 4 D, intensity is weaker in mkt1Δ, its difficult to see the granules. It looks like based on the image that there are less granules, but the quantification shows unchanged granules.
    4. Fig 5 D/ Supl S5A, images with stress are missing for rik1 loss (the ones leading to qualifications in 5D)
    5. Fig 5 C, rik1Δ, ago1Δ, co-colocalised with Cpn1-GFP are missing
    6. Fig.6D, one arrow showing cytoplasmic foci is shifted. PABP stain needs to be added to highlight these are cytoplasmic PABP foci.
    7. cen(dg) transcripts show lack of localization in granules in unstressed cells. The explanation why is a bit unclear. "It remained possible that these foci might rather be associated with RNA degradation, ...". Or RNA degradation happens in the nucleoplasm or cytoplasm. Moreover, in the discussion: "Although we were not able to detect stable accumulation of heterochromatic RNAs in these granules by RNA-FISH, we suspect that this may be because these granules are associated with RNA turnover, although it is also possible that they arise as a result of altered RNP homeostasis more broadly." This is confusing. If those granules form due to accumulation of heterochromatic RNAs, then they cannot be the sites for their degradation, because then they would disassemble, if heterochromatic RNAs are degraded.
    8. Clr4Δ and other conditions shows accumulation of cen(dg) RNA-FISH at the centromeres. It would be informative to see if Cpn1-GFP shows colocalization, which could provide additional evidence for the two models in Fig. 7.
    9. Fig. 6D/E: Its difficult to see changes of cen(dg) RNA-FISH intensity at the centromeres in the images. A close-up view should be provided without overexposure to indicate differences.

    Significance

    This manuscript begins to address a possible relationship between stress granule formation and the regulation of heterochromatin. This is an interesting connection, although the mechanism by which such these two processes are connected is unclear at this time.

  3. Review Commons

    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

    In their manuscript titled "Fission yeast Caprin protein is required for efficient heterochromatin establishment", Zhang and colleagues describe the role of Caprin1 (Cpn1), the fission yeast ortholog of mammalian CAPRIN1, in the de novo establishment of heterochromatin. Using reporter assays for heterochromatin formation, they found that deletion of Cpn1 reduces H3K9 methylation, thereby impairing the de novo establishment of silencing, while the maintenance of silencing at centromeric repeats remains unaffected.

    The authors demonstrate that Cpn1 interacts with the fission yeast orthologs of known human CAPRIN1 interaction partners, Nxt3 and Ubp3. These factors are then tested for their influence on the establishment of silencing.

    Using microscopy to observe tagged Cpn1 and known stress granule markers, the authors show that Cpn1 localizes to stress granules. They also quantitatively assess the impact of Cpn1 interactors on stress granule formation. Additionally, the authors note that different RNAi mutants have stress granules even in the absence of envirmental stresses.

    Finally, they investigate the subcellular localization of the cen(dg) transcript using single-molecule RNA FISH. They find that in heterochromatin mutant cells, cen(dg) transcripts localize to the nucleus and exhibit both nuclear and cytoplasmic localization under glucose starvation conditions. The authors also show, through quantification of RNA FISH and RT-qPCR, that cen(dg) transcripts accumulate in Cpn1 mutants.

    Major comments:

    The authors suggest that Cpn1 is requried for efficient degradation of RNA which in-turn helps the establishemnt of heterochromatin potentially by preventing the formation of R-loops. Yet the localization of Cpn1 under non-stressed conditions is cytoplasmic. Additionally the authors show that Cpn1 helps to limit the accumulation of heterochromatic trancripts on chromatin, yet there is no evidence for a nuclear pool of Cpn1 in these condtions.

    In order to strengthen the link between these nuclear processes it is important that the authors follow up on some of the aspects detailed below.

    Is Cpn1 also present in the nucleus in non-stressed conditions? This would be a pre-requesit for a direct mechanisic link for the model that the authors suggest. This could be, for example, adressed by LeptomycinB treatment followed by imaging of Cpn1. Such an experiment could reveal if the protein is shutteling between the nucleus and the cytoplasm.

    In Fig 6 C the authors show co-localization of dh-transcripts with Papb1 under glucose starvation conditions. To strengthen their hypothesis of cen(dg) RNA binding/regulation by Cpn1 show and quantify co-localization of Cpn1 and cen(dg) transcripts.

    The authors observe cytoplasmic cenRNA in Clr4-delta dhp1-2 cells. To substantiate the hypothesis that Cpn1 binds such transcript co-localization of Cpn1-GFP with cen(dg) transcripts should be examined.

    Can the authors rule out that deletion of Cpn1 affects the RNA levels of proteins important for heterochromatin establishment? As Cpn1 could regulate the stability of mRNAs in the cytoplasm it might be worthwhile to consider also an indirect effect of Cpn1 deletion on the process of heterochromatin establishment. The study would benefit from a genomic characterization of Cpn1-delta cells using RNAseq or an extended discussion of this potential caveat.

    Minor comments:

    RNA-FISH images are labeled cen RNA, please provide consistent lables for the transcript throught the manuscript (cen(dg)).

    Please display individual data points for replicate experiments when displaying qPCR results. This would give the reader more opportunities to judge the distribution of the data points. (Fig2 C,D,F,G,H, Fig6 F)

    Fig 2 G: it is not immediately obvious that the two bar plots display different HOOD amplicons. The presentation of the data could be improved.

    Significance

    In the presented manuscript Zhang and colleauges place Cpn1 as a novel factor into the fission yeast RNAi pathway. This study suggests a link between RNAi and stress granule biology which would provide a novel connection of these fields. The manuscript will be of interest to a specialized audience, both in RNAi/heterochromation formation and stress granule biology and additionally would provide a novel function of a CAPRIN ortholog.

    The manuscript is well written and overall the data is presented well. Furthermore the authors provide a solid genetic characterization of Cpn1's effect on heterochromatin establishment. While the manuscript provides several interesting observations their mechanistic link remains unclear and I belive that it would be very important to substaniate their observations with experiments supporting a direct mechanistic link between heterochromatin establishemnt and Cpn1.

    Field of expertise: small RNA mediated heterochromatin formation, RNA biology, chromatin biology

  4. Review Commons

    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

    Summary:

    In this study, Zhang et al. identified a novel factor named Cpn1 promoting de novo heterochromatin establishment in the fission yeast Schizosaccharomyces pombe. The authors established an elegant genetic approach to identify mutants impaired in the de novo heterochromatin assembly, allowing the quantitative assessment of heterochromatin establishment in a highly reproducible manner. This approach allowed them to revisit potential candidates for heterochromatin establishment from a previous study, leading to the identification of Cpn1. Cells lacking Cpn1 display primarily defects in heterochromatin establishment but not maintenance at constitutive heterochromatin. While the function of Cpn1 was unknown, the authors established a functional link to stress granule formation, demonstrating that Cpn1 is the ortholog of the human RNA-binding protein CAPRIN1, likewise forming a complex with two other factors, Nxt3 and Ubp3. Mutating a putative RNA-binding RRG motif in Cpn1 largely phenocopied the establishment defect seen in the deletion mutant. Moreover, Cpn1 and its complex members co-localize with the stress granule marker poly(A)-binding protein, Pabp. Conversely, deleting Cpn1 or mutating its RRG motif resulted in a reduced number of stress granule formation.

    Providing a further link between heterochromatin and stress granules, the authors showed that heterochromatin-deficient mutants accumulate Papb foci in the absence of stress cells, which was largely dependent on Cpn1. Conversely, the number of stress granules was reduced under stress conditions in these mutants, suggesting that heterochromatic transcripts compete with canonical RNPs that form stress granules. The molecular mechanism by which Cpn1 contributes to heterochromatin establishment remains unclear, though. In contrast to Mkt1, another establishment factor previously studied by the authors, no heterochromatic transcripts were found to be associated with Cpn1 when performing RNA-IP (RIP) experiments. The authors then analyzed pericentromeric transcripts (cen RNA) by smRNA-FISH. Deleting the H3K9 methyltransferase Clr4 resulted in the formation of nuclear foci that co-localized with the centromeric histone variant CENP-A. Glucose starvation increased the number of foci, which were also found in the cytoplasm under this stress condition and partially co-localized with Pabp. Notably, inactivating the exoribonuclease Dph1/Xrn2 also resulted in increased nuclear foci formation and accumulation in the cytosol, which was prevented when Cnp1 was absent. Hence, the authors proposed a model, by which Cpn1 limits accumulation of heterochromatic transcripts on chromatin by facilitating their export and cytoplasmic degradation by Dhp1.

    Major comments

    1. The authors suggest that Cnp1 contributes to heterochromatin establishment by facilitating the removal of excessive heterochromatic transcripts from chromatin. Nevertheless, despite the accumulation of pericentromeric transcripts inside the nucleus in clr4∆, direct evidence for their accumulation on chromatin remains elusive. While the authors cautiously avoid assigning Cnp1 a definite role in heterochromatic transcripts removal, investigating RNA:DNA hybrid accumulation through DNA-RNA immunoprecipitation (DRIP) could strengthen their conclusions. Given the successful application of DRIP by the authors in their Mkt1 study (Taglini et al., 2020) and its prior used by others (PMID: 28404620), this approach appears feasible and judicious when appropriate controls are implemented.
    2. While the data support the hypothesis that Cpn1 binds RNA, the authors could not detect Cpn1 association with heterochromatin transcripts. This could be due to transient interactions and their fast turnover, as the authors suggested. The authors could repeat the RIP experiments in mutants that prevents turnover of transcripts, for instance in dhp1-2 and rrp6∆ mutants.

    Minor comments:

    1. How was the quantification of Pabp and Cpn1 foci performed? Little information is provided ("images were [...] exported to ImageJ analysis"). Given the presence of additional (diffuse) signals even under stress conditions, I'm wondering how foci were distinguished from background? Was there a threshold for signals considered to be 'foci' versus background? The authors should give a more detailed description in the figure legends or Materials & Methods section.
    2. How was the relative sm-FISH intensity in Figure 6D and E determined? Have there been internal controls to ensure that hybridization efficiency was comparable for different strains/samples?

    Significance

    The is an intriguing study that provides several functional links between heterochromatin establishment and stress responses, using a combination of elegant yeast genetics, imaging, biochemical approaches and proteomics. While it was previously shown that heterochromatic transcripts can accumulate on chromatin interfering with heterochromatin assembly (Broenner et al., 2017), this study conceptionally advances our understanding of this process by describing a potential role for Cpn1 in facilitating nuclear export of heterochromatic transcripts. This study further describes the conservation of the human CAPRIN1 complex and its role in cytosolic stress granule assembly in yeast and therefore will be of broad interest for researchers interested in heterochromatin assembly and RNP homeostasis. A limitation of this study is the lack of a distinct molecular mechanism by which Cpn1 promotes heterochromatin establishment. Performing additional experiments could strengthen the authors' arguments and contribute to a better understanding of the underlying mechanism(s).

    I have a longstanding expertise in heterochromatin assembly, transcriptional silencing and yeast genetics using S. pombe and S. cerevisiae as model systems.

  5. Version published to 10.1101/2024.06.19.598224v2 on bioRxiv
  6. Version published to 10.1101/2024.06.19.598224v1 on bioRxiv