Efficiency of heterochromatin establishment is influenced by parental genetic background

To investigate factors influencing heterochromatin establishment, we conducted crosses with different CLRC and/or RNAi mutant strains deficient in heterochromatin, and examined the efficiency of heterochromatin re-establishment in the wild-type progeny. An ade6⁺ reporter gene inserted into the heterochromatic outer repeats of centromere one (cen1:ade6⁺) was employed as a readout: in cells lacking centromeric heterochromatin this reporter is expressed resulting in pale/white colonies, whereas assembly of heterochromatin causes silencing of the reporter resulting in red colonies (Fig 1A). When a reporter-carrying strain lacking heterochromatin (e.g., rik1∆ ago1∆ double mutant, defective for both H3K9 methylation and RNAi), was crossed to a wild-type strain without the reporter, the resulting wild-type (rik1+ ago1+), reporter-carrying progeny were 100% red, consistent with the expected re-establishment of heterochromatin and silencing in this background (Fig 1B). Interestingly, however, when we crossed together two different heterochromatin-deficient strains, e.g., rik1∆ and ago1∆ single mutants, we observed a mix of red and white colonies in the wild-type progeny (~80% red, 20% white). Similar results were seen when crossing other pairs of heterochromatin-deficient strains, although the ratios of red:white colonies varied with parental genotype (S1A Fig). In contrast, the ratio of red:white colonies was highly consistent amongst crosses of different, independent ago1∆ and rik1∆ strains (S1B Fig). It was also consistent regardless of the parent of origin of the reporter (Fig 1B). Passaging of progeny colonies revealed that while the red, silenced state is stable, the white phenotype tends to switch to red over time, consistent with slow heterochromatin re-establishment (Fig 1C). Together these results indicate that, upon reintroduction of missing heterochromatin-related factors by crossing, the efficiency of heterochromatin establishment is dependent on the specific parental background.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Efficiency of heterochromatin re-establishment is influenced by genetic background.

(A) Schematic representation of the cen1:ade6⁺ reporter system. The position of the ade6⁺ insertion at centromere one is shown relative to the centromeric outer repeats (dg and dh), innermost repeats (imr) and central core (cc). Intact heterochromatin is associated with silencing of cen1:ade6⁺ reporter resulting in red colonies on low adenine media; loss of heterochromatin alleviates silencing giving pink/white colonies. (B) Proportions of red (cen1:ade6⁺ silenced) versus white (cen1:ade6⁺ expressed) colonies in the wild-type progeny of crosses between parental strains of the indicated genotypes (underlining denotes strains carrying the cen1:ade6⁺ reporter), based on analysis of n colonies. (C) Passaging of wild-type ‘white’ (WT^W) colonies selected from B. (D) Schematic of the cross-based establishment assay. (E) and (F) Proportions of red (cen1:ade6⁺ silenced) versus pink/white (cen1:ade6⁺ expressed) colonies in the otherwise wild-type progeny of rik1∆ x ago1∆ crosses performed in the indicated genetic backgrounds, based on analysis of n colonies. Relative to wild-type, asterisks denote p ≤ 0.05 (*), p ≤ 0.01 (**) or p ≤ 0.001 (***) from chi-squared (χ²) test analysis.

https://doi.org/10.1371/journal.pgen.1011620.g001

We next tested whether the rate of heterochromatin re-establishment in the progeny resulting from a cross between ago1∆ and rik1∆ strains could be modulated by backgrounds expected to promote or suppress heterochromatin assembly. Since histone deacetylation is required to enable stable H3K9 methylation, we reasoned that the rate of heterochromatin establishment might be increased by interventions that reduce histone H3 acetylation. To test this, we repeated the cross in backgrounds either lacking the histone acetyltransferase Gcn5 [36,37], or over-expressing deacetylases Sir2 and/or Clr3. For example, we deleted gcn5⁺ in ago1∆ and rik1∆ strains, crossed the resulting strains together, and analysed the gcn5∆ ago1+ rik1+ progeny (Fig 1D). As predicted, we found that removal of Gcn5 was sufficient to increase the efficiency of heterochromatin establishment, resulting in the proportion of red colonies in the cross progeny increasing to almost 100% (Fig 1E). Over-expression of Sir2 and/or Clr3 also resulted in a smaller but still significant increase in red colony proportion, consistent with their reported roles in promoting heterochromatin establishment [32,33]. Conversely, we also tested the impact of deleting two factors previously shown to promote heterochromatin establishment, Mkt1 and Tri1 [34,35]. As expected, in the absence of either of these factors the efficiency of heterochromatin establishment appeared reduced, with tri1∆ progeny showing a strong reduction in the proportion of red colonies, and mkt1∆ progeny showing a mixture of white and pink but no red colonies (Fig 1F). Thus, this cross-based system can serve as a sensitized assay to detect and quantify changes in efficiency of heterochromatin establishment.

Cpn1 is a novel factor required for efficient heterochromatin establishment

As a starting point to uncover potential novel heterochromatin establishment factors, we returned to data from a previous genetic screen for factors influencing heterochromatin maintenance [38]. In that screen, maintenance of silencing of the cen1:ade6⁺ reporter was primarily assessed via colony growth on media lacking adenine, which was scored semi-quantitatively on a scale of one to four, where four indicated expression comparable to wild-type, and one indicated strong de-repression. 75 mutant strains were given a growth score of three, suggesting possible mild de-repression of the reporter, but falling below the threshold (of two) set for further follow-up analysis as a candidate maintenance factor. We reasoned that these strains showing potential slight defects in heterochromatin maintenance might include deletions of factors that play a greater role in heterochromatin establishment. To further narrow down the list we selected factors with no known S. cerevisiae ortholog (according to PomBase [39]), since H3K9 methylation and RNAi are absent in this species. This resulted in a shortlist of 13 genes, one of which was known establishment factor Tri1, validating the approach. Of the remaining 12 genes, three were previously uncharacterised, and of these one in particular stood out: SPAC12G12.07c, hereon named cpn1⁺ for reasons outlined below. Like another known establishment factor, Mkt1, the product of this gene has previously been reported to physically associate with components of the MTREC exosome adaptor complex that is implicated in maintenance of facultative heterochromatin domains [40], making it a strong candidate as a novel pathway component. To test this, we assessed the impact of cpn1⁺ deletion in our rik1∆ x ago1∆ establishment assay. Strikingly, cpn1∆ progeny showed a clear reduction in the proportion of red colonies, similar to what was seen in the tri1∆ background, suggesting that the product of this gene is required for efficient heterochromatin establishment (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Cpn1 is a novel factor required for efficient heterochromatin establishment.

(A) Assay for establishment of silencing of the cen1:ade6⁺ reporter. Shown are proportions of red (ade6⁺ silenced) versus white (ade6⁺ expressed) colonies in the otherwise wild-type progeny of rik1∆ x ago1∆ cross performed in wild-type or cpn1∆ background (as in 1D), based on analysis of n colonies. Asterisks denote p ≤ 0.001 (***) from chi-squared (χ²) test analysis. (B) Assay for maintenance of silencing of the cen1:ade6⁺ reporter (red colonies indicate silencing; white colonies loss of silencing), and for sensitivity to TBZ (20 μg/μl). (C) RT-qPCR analysis of cen(dg) transcript levels relative to act1⁺, normalised to wild-type. (D) ChIP-qPCR analysis of H3K9me2 levels at cen(dg) relative to act1⁺, normalised to wild-type. (E) Schematic of the minichromosome (MC-dg) plasmid. (F) ChIP-qPCR analysis of H3K9me2 levels established de novo on the MC-dg plasmid, relative to act1⁺, normalised to wild-type. (G) ChIP-qPCR analysis of H3K9me2 levels at HOOD locus SPCC1442.04c, relative to act1⁺, normalised to wild-type. (H) ChIP-qPCR analysis of H3K9me2 levels at cen(dg) relative to act1⁺, normalised to wild-type. In each case data are averages of three biological replicates; dots represent individual data points and error bars represent one SD. Asterisks denote p ≤ 0.05 (*), p ≤ 0.01 (**) or p ≤ 0.001 (***) from Student’s t-test analysis.

https://doi.org/10.1371/journal.pgen.1011620.g002

Our previous screen results had suggested that deletion of cpn1⁺ might result in a small defect in heterochromatin maintenance [38]. To assess this in more detail, we deleted cpn1⁺ in a wild-type strain carrying the cen1:ade6⁺ reporter to assess the effect on heterochromatin maintenance. Colony colour of cpn1∆ strains was indistinguishable from wild-type, suggesting little or no effect on maintenance of reporter gene silencing (Fig 2B). Consistent with this, cpn1∆ strains, in contrast to clr4∆ strains, showed little increase in sensitivity to the microtubule destabilising drug TBZ, indicating that centromere function is maintained. This was further supported by RT-qPCR analysis that showed no significant increase in accumulation of non-coding pericentromeric (dg) transcripts (Fig 2C). In addition, ChIP-qPCR analysis of H3K9me2 revealed only a small reduction in H3K9 methylation levels at pericentromeric repeat sequences (Fig 2D). Hence loss of Cpn1 function has little effect on heterochromatin maintenance, and rather appears to particularly affect heterochromatin establishment.

As a complementary approach to further verify and characterise the heterochromatin establishment defects in cpn1∆ cells, we employed an alternative assay whereby we assessed de novo heterochromatin establishment on a newly introduced minichromosome, as described previously [32,35]. Wild-type and mutant strains were transformed with a minichromosome plasmid (MC-dg) that carries a 5.6kb portion of centromeric (dg) outer-repeat sequence that is targeted by endogenous centromeric siRNAs (Fig 2E). ChIP-qPCR analysis revealed that, whereas in wild-type cells, heterochromatin is efficiently established on the plasmid resulting in high levels of H3K9me2, in cells lacking Cpn1 the levels of H3K9me2 established on the plasmid are substantially reduced (Fig 2F). This is consistent with the reduced efficiency of re-initiation of pericentromeric reporter gene silencing observed in cpn1∆ cells, and supports the conclusion that absence of Cpn1 is associated with defective heterochromatin establishment.

Cpn1 has previously been reported to interact with MTREC components including Red1, Red5, Mtl1 and Pir2 [40]. We therefore tested whether Cpn1 also functions together with MTREC in facilitating assembly of facultative heterochromatin. Two classes of MTREC-dependent facultative heterochromatin domains can be distinguished: heterochromatin islands, which are RNAi-independent and typically serve to silence meiotic genes during vegetative growth [17,18], and so-called HOODs, which are dependent on both MTREC and RNAi and form at loci such as transposons in the absence of Rrp6 [19]. In the case of heterochromatin islands, RT-qPCR analysis revealed that, while absence of MTREC component Red1 results in loss of silencing of meiotic genes such as mei4⁺ and ssm4⁺, the removal of Cpn1 has no effect at these loci (S2A Fig). Similarly, at HOODs (assessed in rrp6∆ background), ChIP-qPCR analysis indicated that, whereas H3K9 methylation is lost upon deletion of dcr1⁺, it is retained at most loci tested in the absence of cpn1⁺ (S2B Fig). However, the one exception to this is the HOOD associated with meiotic gene SPCC1442.04c, where we found heterochromatin to be strongly dependent on functional Cpn1 (Fig 2G). This was striking since we previously found that this particular HOOD (but not other HOODs or islands) is also dependent on Mkt1 (but not Tri1) [35]. This HOOD is also unusual in being Red1-independent [19]. Hence Cpn1, like Mkt1, plays a role in supporting RNAi-dependent facultative heterochromatin formation in certain contexts, possibly acting in parallel to Red1.

The commonalities described above prompted us to probe whether Mkt1 and Cpn1 might be functionally linked. We showed previously that Mkt1 acts in the same pathway as RNAi for pericentromeric silencing. To test if this is also true for Cpn1, we assessed pericentromeric H3K9me2 levels in cpn1∆ ago1∆ double mutant cells. In contrast to our previous observations for Mkt1, we found that the cpn1∆ ago1∆ double mutants displayed additive defects in H3K9 methylation as compared to either single mutant, suggesting independent effects (Fig 2H). Hence it appears that while Mkt1 functions together with RNAi, Cpn1 functions in another, separate pathway.

Cpn1 is a CAPRIN family protein

At the time of this study no orthologs of the SPAC12G12.07c gene product had been identified outside of fungi. However, through an iterative search for remote homology using JACKHMMER [41], we identified 1:1 orthology with human protein CAPRIN1. Although they have only 15% identity at the amino level, the human and fission yeast proteins share similar motif organisation, including N-terminal coiled-coil and C-terminal RG/RGG-rich regions (Fig 3A). Moreover, comparison of structural models generated by AlphaFold [42,43] also reveals clear similarities in structural organisation (Fig 3B). We therefore named the SPAC12G12.07c gene product as fission yeast Caprin protein, Cpn1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Cpn1 is a CAPRIN family protein.

(A) Schematic representation of CAPRIN proteins in human (CAPRIN1) and S. pombe (Cpn1). (B) Alphafold structural models of Cpn1 and CAPRIN1. (C) List of proteins found specifically and reproducibly associated with Cpn1-FLAG by affinity purification and mass spectrometry (LC-MS/MS). Shown are total peptide counts for all proteins represented by ≥ 2 peptides in purifications from Cpn1-FLAG strains, but not untagged control strains, in two independent experiments. (D) Putative Cpn1 complex (dotted lines denote interactions detected here; solid line indicates interaction detected previously by IP-western [45]). (E) Assay for maintenance of silencing of cen1:ade6⁺ reporter: red colonies on low adenine media indicate silencing, and white colonies absence of silencing. (F) Assay for establishment of silencing of cen1:ade6⁺ reporter: shown are proportions of red (ade6⁺ silenced) versus white (ade6⁺ expressed) colonies in the otherwise wild-type progeny of rik1∆ x ago1∆ crosses performed in the indicated deletion backgrounds (as in 1D), based on analysis of n colonies. Relative to wild-type, asterisks denote p ≤ 0.001 (***) from chi-squared (χ²) test analysis. (G) Western blot analysis of affinity-purified FLAG-tagged wild-type Cpn1 and Cpn1^R6A.

https://doi.org/10.1371/journal.pgen.1011620.g003

Human CAPRIN1 is a ubiquitously expressed mRNA-binding protein that is known to be involved in regulation of stress granule formation in association with two partner proteins, G3BP1 and USP10 [44]. To investigate fission yeast Cpn1 interaction partners, we C-terminally FLAG-tagged Cpn1 at the endogenous locus and performed immunoprecipitation followed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) to identify associated proteins. Only five proteins were specifically and reproducibly identified in precipitates from Cpn1-FLAG-expressing cells but not controls, and strikingly, two of these were Nxt3 and Ubp3, the fission yeast orthologs of human CAPRIN1 partner proteins G3BP1 and USP10, respectively (Fig 3C and 3D). Nxt3 and Ubp3 are themselves known to interact, and to localise to stress granules [45,46]. The association with Nxt3 and Ubp3 further supports the conclusion that fission yeast Cpn1 is the functional ortholog of human CAPRIN1.

To explore whether Nxt3 and Ubp3 also function together with Cpn1 to influence heterochromatin assembly, we generated nxt3∆ and ubp3∆ deletion strains and tested them in heterochromatin maintenance and establishment assays. While absence of these partner proteins, like loss of Cpn1 itself, had little effect on heterochromatin maintenance (Fig 3E), it did affect heterochromatin establishment in our rik1∆ x ago1∆. Interestingly, the two deletions had opposing effects: while in nxt3∆ cells, as in cpn1∆ cells, the efficiency of heterochromatin establishment was reduced (lower proportion of red colonies in the cross progeny compared to wild-type), in ubp3∆ cells we conversely saw a small increase in establishment efficiency (increased proportion of red colonies; Fig 3F). Considering what is known about their human orthologs, this may be explained by opposing functions of these proteins, since CAPRIN1 and G3BP1 are both RNA-binding proteins that have been shown act cooperatively to promote stress granule assembly, while USP10, which lacks an RNA-binding domain, acts to limit assembly [44]. Hence Cpn1 acts together with its interaction partners to influence heterochromatin establishment efficiency.

The RGG-rich region in human CAPRIN1 has been shown to be important for RNA-binding and stress granule formation [47]. To assess whether the RG/RGG-rich motif in fission yeast Cpn1 is also required for its function in heterochromatin establishment, we generated strains in which each of the six arginine residues in this region were mutated to alanine (Cpn1^R6A; Fig 3A). Testing this mutant in our rik1∆ x ago1∆ establishment assay, we observed a reduction in the proportion of red colonies in the progeny compared to wild-type, indicating reduced efficiency of heterochromatin establishment, although the effect was not as strong as that associated with cpn1⁺ deletion (Fig 3F). Moreover, in Cpn1^R6A-expressing cells we also observed loss of H3K9me2 from the SPCC1442.04c HOOD locus, and a reduction in H3K9me2 maintenance at pericentromeric repeats in ago1∆ background, comparable to that in cpn1∆ cells (Fig 2G and 2H). Western blot analysis of FLAG-tagged wild-type and mutant Cpn1 proteins showed no difference in protein accumulation (Fig 3G), confirming that the mutations do not affect protein stability. Hence the RGG-rich region contributes to Cpn1 function in facilitating heterochromatin assembly.

Since the RGG motifs in human CAPRIN1 are involved in RNA binding, and since we have shown previously that Mkt1 interacts with non-coding pericentromeric transcripts [35], we tested whether Cpn1 might also bind to pericentromeric RNAs. As previously, we performed RNA-IPs in cells lacking Sir2, to relieve transcriptional repression at pericentromeric heterochromatin [35]. We detected specific enrichment of pericentromeric transcripts (dg and imr) in Mkt1-FLAG pulldowns relative to controls, consistent with our previous results. However, no such enrichment was observed in Cpn1-FLAG pull-downs, suggesting that Cpn1 does not stably bind these RNAs (S3 Fig). We also tested whether association of Mkt1 with pericentromeric transcripts is dependent on Cpn1, but again found no evidence of this, with both dg and imr transcripts being similarly enriched in Mkt1-FLAG pull downs in the presence and absence Cpn1 (S3 Fig). This supports the conclusion that Cpn1 and Mkt1 function independently in heterochromatic silencing.

To explore possible wider effects of Cpn1 on the transcriptome, and rule-out the possibility that loss of Cpn1 might affect heterochromatin establishment indirectly via altered expression of one or more heterochromatin pathway components, we performed transcriptome analysis by RNA sequencing in cpn1∆ cells. Very few transcripts were found to be differentially expressed in cpn1∆ cells compared to wildtype (11 up-regulated, 15 down-regulated; S1 and S2 Tables, respectively). Importantly, no genes with known links to heterochromatin assembly showed altered expression, arguing against a possible indirect effect of Cpn1 removal. Interestingly, the majority of upregulated transcripts were non-coding RNAs, including pericentromeric transcripts, as well as transcripts from a small number of other, non-heterochromatic loci, consistent with a role for Cpn1 in non-coding RNA regulation. Moreover, nearly all of the down-regulated transcripts derived from subtelomeric loci (within 100kb of the ends of chromosomes I or II). Given that Cpn1 impacts centromeric heterochromatin assembly, and that subtelomeres are also heterochromatic, we hypothesised that this down-regulation might reflect increased heterochromatinisation in these regions. Consistent with this, ChIP-qPCR analyses revealed a trend towards elevated levels of H3K9 methylation at subtelomeric loci in cpn1∆ cells (S4 Fig). Increased H3K9 methylation was most clearly detected at sites just outside the normal heterochromatin domain boundaries seen in wild-type cells [48], indicating heterochromatin spreading. Previous studies have indicated that subtelomeric heterochromatin domains have flexible boundaries and can serve as sinks for excess heterochromatin factors [49]; we therefore suspect that redistribution of silencing factors indirectly accounts for the increased repression of subtelomeric genes in cpn1∆ cells, although we also cannot rule out a more direct role for Cpn1 in regulation at subtelomeres.

Cpn1 promotes stress granule formation

To test whether Cpn1, like its human ortholog, localises to stress granules, we introduced a C-terminal GFP tag at the endogenous locus and performed live cell imaging analyses. While in unstressed cells we saw a diffuse pattern of localisation, in cells subjected to stress (heat shock or glucose starvation), Cpn1 could be seen to accumulate in discrete cytoplasmic foci. Analysis of cells co-expressing stress granule marker poly(A)-binding protein Pabp tagged with RFP [50] indicated that Cpn1 colocalises with Pabp, confirming that these foci correspond to stress granules (Fig 4A). Similar results were obtained for GFP-tagged Cpn1^R6A, indicating that the RGG-rich region is not required for Cpn1 localisation to stress granules. We also generated strains expressing C-terminally GFP-tagged Nxt3 or Ubp3, and confirmed that both of these proteins also colocalise with both Pabp and Cpn1 in stress-induced granules (Fig 4B and 4C). Western blot and heterochromatin establishment assays confirmed that all of the GFP-tagged proteins are stably expressed and functional (S5 Fig). Hence Cpn1, together with its partner proteins Nxt3 and Ubp3, does indeed partition into phase-separated, membrane-less organelles in response to stress.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Cpn1 localises to stress granules and promotes their formation.

(A) Representative images from two-colour live-cell imaging of Cpn1-GFP or Cpn1^R6A-GFP (green), and Pabp-mRFP (red), in cells grown at 32^oC (control) and exposed to 42^oC heat shock for 20 min, or glucose starvation for 20 min. Bar indicates 6 µm. (B) and (C) Representative images from two-colour live-cell imaging of Nxt3-GFP or Ubp3-GFP (green), and either Pabp-mRFP (red; B) or Cpn1-mCherry (red; C), in cells either untreated or treated with heat shock as above. (D) Representative images from live-cell imaging of Pabp-mRFP (red) in the indicated genetic backgrounds in cells exposed to heat shock as above. (E) Quantification of the mean number of Pabp-mRFP foci per cell from the imaging analysis shown in D (n = 100). Error bars represent one SD and asterisks denote p ≤ 0.001 (***) from Student’s t-test analysis.

https://doi.org/10.1371/journal.pgen.1011620.g004

Human CAPRIN1 has also been shown to be required to promote stress granule assembly, in a manner dependent on its RGG repeats [44,47]. To test if the same is true for fission yeast Cpn1, we quantified Pabp focus formation in response to stress in cells lacking Cpn1 or other factors. In line with predictions, in cells lacking Cpn1 we found that the average number of Pabp foci formed per cell following heat shock was substantially reduced (Fig 4D and 4E). Granule formation was also reduced in cells expressing Cpn1^R6A, suggesting that the RGG-rich motif is important for Cpn1 function in stress granule assembly. We also observed a reduction in granules in cells lacking Nxt3, consistent with findings for the human ortholog G3BP1 [44]. In contrast, no change in Pabp focus formation was observed in cells lacking Ubp3, or Mkt1 (Fig 4D and 4E). Hence like their human orthologs, Cpn1 and Nxt3 are required to promote stress granule formation.

Interplay between heterochromatin integrity and cytoplasmic RNP granule formation

We next set out to explore the possible connection between assembly of heterochromatin and RNP granules. First, we tested whether mutations that disrupt heterochromatin impact on granule formation. Interestingly, whereas in the wild-type background, Pabp foci are not seen in unstressed cells, when we deleted genes encoding components of either the RNAi machinery (Ago1) or the CLRC H3K9 methyltransferase complex (Clr4 or Rik1), we observed a proportion of cells exhibiting Pabp foci even in the absence of stress (Fig 5A and 5B). This was most prominent in ago1∆ cells, where granules were visible in approximately 15% of cells. Analysis of cells co-expressing Cpn1-GFP confirmed that Cpn1 also localises to these granules, as for canonical stress granules (Fig 5C; colocalization was seen in all cells displaying visible granules). We further tested whether the formation of these ago1∆ or clr4∆-dependent granules is dependent on Cpn1. Indeed, we found that granule formation was greatly reduced in cpn1∆ ago1∆ or cpn1∆ clr4∆ double mutants compared to the respective single mutants, indicating that Cpn1 is required to promote the formation of these heterochromatin deficiency-induced granules (Fig 5B). Western blot analysis indicated that Cpn1 protein levels are unaffected by absence of Clr4, Rik1, or Ago1, confirming that the appearance of granules in heterochromatin mutants is not due to altered Cpn1 expression (S6 Fig). Similarly, previous expression analyses have shown that loss of Clr4 or Ago1 has little overall effect on the transcriptome in standard conditions [51–54], arguing against the possibility that increased granule formation reflects a global change in mRNA abundance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Disruption of heterochromatin is associated with altered formation of Pabp-containing RNP granules.

(A) Representative images from live-cell imaging of Pabp-mRFP (red) in the indicated genetic backgrounds in unstressed cells. Bar indicates 6 µm. (B) Quantification of the proportion of unstressed cells displaying Pabp-mRFP foci from the imaging analysis shown in A, based on analysis of n cells. Asterisks denote p ≤ 0.001 (***) from chi-squared (χ²) test analysis. (C) Representative images from two-colour live-cell imaging of Cpn1-GFP (green), and Pabp-mRFP (red), in the indicated genetic backgrounds in unstressed cells. (D) Representative images from live-cell imaging of Pabp-mRFP (red) in the indicated genetic backgrounds in cells exposed to 42^oC heat shock for 20 min. (E) Quantification of the mean number of Pabp-mRFP foci per cell in the indicated genetic backgrounds in unstressed cells (based on analysis of n cells containing foci from the imaging analysis shown in A), or cells exposed to 42^oC heat shock for 20 min (n = 100 from the imaging analysis shown in D; virtually all cells contained foci). Error bars represent one SD and asterisks denote p ≤ 0.001 (***) from Student’s t-test analysis.

https://doi.org/10.1371/journal.pgen.1011620.g005

We also tested what happens in heterochromatin-deficient cells in the presence of stress. Interestingly, we found that, 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 response to heat shock, compared to wild-type cells (Fig 5D). Analysis of foci area confirmed that mean focus size was unchanged in the mutants, but mean foci area per cell reduced, indicating reduced stress granule formation in these backgrounds (S7A Fig). Surprisingly, the same effect was not seen in cells subject to glucose starvation, suggesting differences in the responses to these two forms of stress (S7B and S7C Fig). Together, these observations suggest a degree of interplay between heterochromatin integrity and stress granule formation that could reflect involvement of one or more common limiting factors.

Cpn1 colocalises with heterochromatic transcripts

We hypothesised that the RNP granules that are occasionally formed in response to heterochromatin disruption may be linked to sequestration and/or degradation of excess heterochromatic transcripts that accumulate in the absence of heterochromatin-associated silencing. While previous fractionation analyses have provided evidence that heterochromatic transcripts tend to be retained on chromatin [24], to our knowledge there has been no direct visualisation of heterochromatic transcript subcellular localisation in S. pombe. To explore whether the RNP granules formed in response to heterochromatin disruption might harbour non-coding pericentromeric RNAs, we therefore set out determine the subcellular localisation of these RNAs by single‐molecule mRNA fluorescence in situ hybridization (smRNA-FISH) using probes targeting cen(dg) transcripts. Previous RT-qPCR and northern blot analyses have indicated that pericentromeric transcript levels are low in wild-type cells, but greatly elevated in heterochromatin mutants where transcriptional silencing is alleviated [6,10,55,56]. Consistent with this, by smRNA-FISH we could detect a strong signal for cen(dg) transcripts in clr4∆ cells, but not in wild-type cells, confirming that the signal is cen(dg) transcript-specific (Fig 6A). Notably, cen(dg) transcripts were detected in one strong focus in the nucleus, which we suspected might corresponded to the position of centromeres, which tend to cluster together at the nuclear periphery in S. pombe [57]. To test this, we combined smRNA-FISH with analysis of GFP-tagged centromeric histone variant CENP-A^Cnp1. The cen(dg) transcript focus was consistently found adjacent to, or overlapping, the CENP-A^Cnp1 focus, consistent with transcripts being primarily retained at the pericentromeres (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Cpn1 colcalises with heterochromatic transcripts in heterochromatin deficient cells.

(A) Representative images from smRNA-FISH analysis of pericentromeric (cen(dg)) RNA in wild-type or clr4∆ strains. Data are representative of three independent experiments. Bar indicates 6 µm. (B) Representative images from simultaneous analysis of cen(dg) RNA by smRNA-FISH, and GFP-CENP-A^Cnp1, in clr4∆ cells. Data are representative of two independent experiments. (C) and (D) Representative images from simultaneous analysis of cen(dg) RNA by smRNA-FISH, and Pabp-GFP (C), or Cpn1-GFP (D), in clr4∆ cells either unstressed (+ glucose) or stressed by 20 min of glucose starvation (- glucose). Arrow heads highlight examples of co-localisation (Pearson’s correlation coefficient, R, is indicated, and inserts show 5x magnification). Data are representative of at least three independent experiments.

https://doi.org/10.1371/journal.pgen.1011620.g006

To address whether cen(dg) transcripts might also localise to the Pabp foci seen in a small proportion of (unstressed) heterochromatin-deficient cells, we combined RNA-FISH with simultaneous analysis of Pabp-GFP. In those clr4∆ cells where visible Pabp foci were formed, we detected no additional cen(dg) transcript foci, and hence no co-localisation of these transcripts with Pabp foci in the cytoplasm (Fig 6C, upper panel). This suggests that cytoplasmic Pabp foci formed in this background are not sites of stable heterochromatic transcript accumulation. However, in almost half of clr4∆ cells with Pabp foci, a nuclear Pabp focus could be observed, and this nuclear Pabp signal frequently co-localised with the cen(dg) transcript focus (Pearson’s correlation coefficient, R_p = 0.82, for the 43% of cells displaying nuclear Pabp foci). Nuclear colocalisation of Pabp with cen(dg) transcripts was also observed in a proportion of ago1∆ cells (S8A Fig, upper panel; R_p = 0.85, for the 24% of cells displaying nuclear Pabp foci). Moreover, visualisation of Cpn1-GFP in clr4∆ cells similarly revealed some localisation of Cpn1 in the nucleus (in 35% of cells displaying Cpn1-GFP foci), with the nuclear Cpn1-GFP focus also often colocalising with cen(dg) transcripts (Fig 6D upper panel; R_p = 0.80). Hence, disruption of heterochromatin is associated with some accumulation of the predominantly cytoplasmic Cpn1 protein in the nucleus; this observation is consistent with a previous study that showed accumulation of Cpn1 in the nucleus following treatment of cells with leptomycin B, which inhibits CRM1-dependent nuclear export [58]. That Cpn1 can enter the nucleus and co-localise with accumulating heterochromatic transcripts is consistent with it playing a direct role in their regulation.

Although the cytoplasmic Pabp/Cpn1 foci occasionally seen in unstressed, heterochromatin-deficient cells appeared not to be sites of stable cen(dg) transcript accumulation, it remained possible that heterochromatic transcripts might be exported to the cytoplasm, possibly for degradation. Moreover, co-localisation of Cpn1 and Pabp with heterochromatic transcripts in the nucleus of unstressed cells suggested that these transcripts might be expected to show altered localisation in conditions of stress, where Pabp and Cpn1 consistently relocalise into cytoplasmic granules. To explore these possibilities, we performed additional RNA-FISH analyses in clr4∆ or ago1∆ cells subject to stress. Notably, following glucose starvation, additional cen(dg) transcript foci were observed in the cytoplasm, at least some of which colocalised with Pabp-GFP-marked stress granules (Figs 6C and S8A, lower panels, R_p = 0.42 and 0.40, respectively). The stress-induced cytoplasmic cen(dg) transcript foci also showed co-localisation with Cpn1-GFP (Fig 6D, lower panel, R_p = 0.42). No DAPI signal was detected in association with these cytoplasmic cen(dg) transcript foci, arguing against the possibility that they arise as a result of micronuclei formation. Hence these observations indicate that heterochromatic transcripts can exit the nucleus, and co-localise with Cpn1 and associated factors in the cytoplasm when these proteins relocalise during stress.

Cpn1 helps limit accumulation of heterochromatic transcripts on chromatin

To assess whether Cpn1 impacts pericentromeric transcript localisation and/or accumulation in the absence of stress, we performed further RNA-FISH analyses in cells lacking Cpn1. Interestingly, we observed a visible increase in the intensity of the cen(dg) transcript signal at centromeres in clr4∆ cpn1∆ double mutants compared to clr4∆ single mutants, suggesting that Cpn1 is required to help limit accumulation of these transcripts at centromeres (Fig 7A and 7B). The same effect was also seen in ago1∆ background (S8B and S8C Fig). DNA-RNA immunoprecipitation (DRIP) analyses also revealed an increase in DNA-RNA hybrid accumulation at centromeres in clr4∆ cpn1∆ double mutants compared to clr4∆ single mutants, consistent with excess RNAs accumulating on chromatin (Fig 7D). In principle, the Cpn1-dependent protection of centromeres from excess RNA accumulation might be achieved either via sequestration of these RNAs elsewhere in the cell, or through their degradation. To distinguish between these two possibilities, we analysed total cellular levels of cen(dg) transcripts by RT-qPCR. Strikingly, this revealed that deletion of cpn1⁺ in either clr4∆ or ago1∆ background results in a significant increase in total cen(dg) transcripts levels, beyond the already high levels present in absence of Clr4 or Ago1 alone (Figs 7C and S8D). Thus, Cpn1 appears to be required to facilitate degradation of excess heterochromatic transcripts, limiting their accumulation on chromatin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Absence of Cpn1 results in increased accumulation of heterochromatic transcripts at centromeres in heterochromatin deficient cells.

(A) Representative images from smRNA-FISH analysis of cen(dg) RNA in clr4∆, clr4∆ cpn1∆, clr4∆ dhp1-2, and clr4∆ cpn1∆ dhp1-2 cells. Arrow heads highlight examples of cytoplasmic RNA foci; Bar indicates 6 µm. Data are representative of at least three independent experiments. (B) Quantification of the mean signal intensity for nuclear cen(dg) RNA foci, normalised to clr4∆ cells, from the smRNA-FISH analysis shown in A. (C) RT-qPCR analysis of total cellular levels of cen(dg) RNA relative to act1⁺, normalised to clr4∆ cells. (D) DRIP analysis of DNA:RNA hybrid levels at cen(dg) relative to act1⁺, normalised to clr4∆ cells. Data are averages of three biological replicates, dots represent individual data points, and error bars represent one SD. Asterisks denote p ≤ 0.05 (*), or p ≤ 0.001 (***), from Student’s t-test analysis. (E) Proportions of red (cen1:ade6⁺ silenced) versus white (cen1:ade6⁺ expressed) colonies in the otherwise wild-type progeny of rik1∆ x ago1∆ crosses performed in the indicated genetic backgrounds, based on analysis of n colonies. Asterisks denote p ≤ 0.001 (***) from chi-squared (χ²) test analysis. (F) Representative images from simultaneous analysis of cen(dg) RNA by smRNA-FISH, and either Pabp-GFP or Cpn1-GFP, in clr4∆ dhp1-2 cells. Arrow heads highlight examples of co-localisation (Pearson’s correlation coefficient, R, is indicated, and inserts show 5x magnification). Data are representative of two independent experiments.

https://doi.org/10.1371/journal.pgen.1011620.g007

Finally, we reasoned that if Cpn1 facilitates RNA degradation in the cytoplasm, suppression of degradation (independently of other Cpn1 functions) might lead to visible cytoplasmic accumulation of cen(dg) transcripts. Since Caprin proteins do not themselves appear to have nuclease activity, it is likely that Cpn1 primarily influences transcript localisation and/or accessibility for degradation by one or more ribonucleases. We identified the 5’-3’ exoribonuclease Dhp1 (ortholog of human Xrn2) as a likely candidate, since: (1) in human embryonic stem cells, Xrn2 has been implicated in mediating CAPRIN1-dependent degradation of selected transcripts in the cytoplasm [59]; and (2) Dhp1, like Cpn1, has been shown to be required for efficient heterochromatin establishment in S. pombe [22, 23]. We confirmed that Dhp1 is required for heterochromatin establishment in our system: because dhp1⁺ is an essential gene, we tested the effects of a previously-described temperature-sensitive allele, dhp1-2, that causes defects in heterochromatin, but little impairment of growth, at the permissive temperature [23]. Performing our rik1∆ x ago1∆ establishment assay in the dhp1-2 mutant background resulted in a reduced proportion of red colonies in the progeny compared to wild-type, similar to what we saw for cpn1⁺ deletion, consistent with the possibility that these factors might act together in heterochromatin establishment (Fig 7E). To explore the influence of Dhp1 on pericentromeric transcript accumulation, we performed further RNA-FISH analyses in dhp1-2 mutant cells. Strikingly, whereas in clr4∆ cells a single cen(dg) transcript focus can be seen in the nucleus, in dhp1-2 clr4∆ cells, we observed the appearance of additional cen(dg) RNA foci in the cytoplasm (Fig 7A; cytoplasmic RNA foci were detected in ~22% of dhp1-2 clr4∆ cells, possibly reflecting the fact that dhp1-2 is a hypomorphic allele). This indicates that active degradation prevents accumulation of cen(dg) transcripts in the cytoplasm in clr4∆ cells in normal conditions. Importantly, the cytoplasmic cen(dg) RNA foci were abolished in cpn1∆ dhp1-2 clr4∆ triple mutant cells, consistent with loss of Cpn1 causing a defect upstream of Dhp1, likely in the localisation of transcripts for degradation (Fig 7A). Moreover, we confirmed that a proportion of the cytoplasmic cen(dg) RNA foci that appear in dhp1-2 clr4∆ cells show colocalisation with Pabp-GFP, and with Cpn1-GFP (Fig 7F; Pearson’s Correlation Coefficients 0.47 and 0.46, respectively), further supporting their functional association. Together, our data suggest that Cpn1 helps to enable heterochromatic transcript degradation by Dhp1, limiting excess accumulation of RNA on chromatin that may otherwise interfere with heterochromatin assembly.

Yeast strains and methods

Strains used in this study are listed in S3 Table. Standard procedures were used for growth and genetic manipulations. Cultures were grown at 32^oC unless otherwise stated, in either rich medium (YES) or minimal media (PMG). Genomic integrations for gene deletion and epitope-tagging were achieved by homologous recombination using PCR-based modules consisting of a resistance cassette flanked by 80 bp sequences homologous to the target locus. To generate the cpn1^R6A mutant, the six arginines in the RGG motif were mutated to alanine by CRISPR/Cas9 editing as previously described [77]. Genomic modifications were verified by sequencing. Overexpression plasmids pSP102-Sir2, pSP102-Clr3 and pSP102-Sir2-Clr3 were described previously [78]. The minichromosome establishment assay was carried out as described previously [32]. For cross-based establishment assays, random spore analysis was employed. Strains bearing different gene deletions marked with ura4⁺ were crossed, and the products plated onto selective media containing 1g/L 5-FOA and low (1/10^th) adenine to directly select for desired recombinant genotypes (and against parental genotypes) prior to counting of red/white colonies.

Live cell imaging

Visualisation of fluorescently-tagged proteins in living cells was performed on log phase cultures grown at 32°C in PMG and, where indicated, pre-treated as follows: heat shock, 20 min incubation at 42°C; glucose starvation, 20 min incubation in medium lacking glucose. Images were acquired using a Nikon Ti2 inverted microscope equipped with a 100x 1.49 NA Apo TIRF objective and a Teledyne Photometrics Prime 95B camera. Images were acquired with NIS-elements (version 5.1), with 25 z-stacks taken at 0.2µm intervals, deconvolved with AutoQuant X3 software and subsequently exported to ImageJ for analysis. Quantification of foci number/area was achieved by intensity threshold segmentation followed by application of the ‘analyse particles’ function of ImageJ.

smRNA FISH

smRNA FISH was performed essentially as described previously [79]. Briefly, 2×10⁸ cells were fixed with 4% paraformaldehyde for 30 min, or 2% paraformaldehyde for 30 min for simultaneous imaging of GFP-tagged Pabp or Cpn1. Cells were washed with Buffer B (1.2 M sorbitol, 0.1M potassium phosphate buffer pH 7.5), resuspended in spheroplast buffer (1.2 M sorbitol, 0.1 M potassium phosphate, 20 mM vanadyl ribonuclease complex [NEB, S1402S], 20 μM beta-mercaptoethanol) and digested with 0.002% 100 T zymolyase (US Biological, Z1005) for approximately 45–75 min. Cells were washed with Buffer B, resuspended in 1 ml of 0.01% Triton X-100 in 1x PBS for 20 min, washed again with Buffer B, and resuspended in 10% formamide/2x SSC buffer. For hybridization of probes, approximately 20–25ng of CAL Fluor red 610 probes targeting cen(dg) repeat sequences (Stellaris, Biosearch Technologies) were mixed with 2 μl each of yeast tRNA (Life Technologies) and Salmon sperm DNA (Life Technologies) per reaction. Sequences of the cen(dg) probes are given in S5 Table. The probe solution was mixed with Buffer F (20% formamide, 10 mM sodium-phosphate buffer pH 7.2; 45 μl per reaction), heated at 95°C for 3 min, and allowed to cool to room temperature before mixing with Buffer H (4x SSC, 4 mg/ml acetylated BSA, 20 mM vanadyl ribonuclease complex; 50 μl per reaction). Digested cell samples were divided into two reactions, each of which was resuspended in 100 μl of this hybridization solution and incubated at 37°C overnight. Cells were subsequently washed sequentially with: 10% formamide/2x SSC; 0.1% Triton X-100/2x SSC; 2x SSC; 1x SSC; and 1x PBS. Cell pellets were mixed with VECTASHIELD Vibrance Antifade Mounting Medium with DAPI (Vector Laboratories) and mounted on poly-L-Lysine slides with 1.5mm glass coverslips. Imaging was performed using a Nikon Ti2 inverted microscope equipped with a 100x 1.49 NA Apo TIRF objective and a Teledyne Photometrics Prime 95B camera. Images were acquired with NIS-elements (version 5.1), with 25 z-stacks taken at 0.2µm intervals, further deconvolved with AutoQuant X3 software and subsequently exported to ImageJ for analysis. Quantification of RNA foci intensity was achieved by intensity threshold segmentation followed by application of the ‘analyse particles’ function of ImageJ.

RT-qPCR

Total RNA was extracted from 1 × 10⁷ cells in exponential growth phase using the Masterpure Yeast RNA Purification Kit (Epicentre) according to the manufacturer’s instructions. 1 μg of total RNA was treated with TURBO DNase (Ambion) for 1 h at 37°C, then reverse transcribed using random hexamers (Roche) and Superscript III reverse transcriptase (Invitrogen) according to the manufacturer’s instructions. cDNA was quantified by qPCR using LightCycler 480 SYBR Green (Roche) and primers listed in S4 Table. In all cases, histograms represent three biological replicates and error bars represent one S.D.

RNA-seq

Total RNA was isolated as above and provided to Azenta Life Sciences for polyA selection, library preparation, and sequencing using an Illumina Novaseq X, generating a minimum of 30 million 150 bp paired-end reads per sample. Three biological replicates for wild-type and cpn1∆ were processed in parallel. Adapters were removed from raw reads using Cutadapt (v4.8) [80] and the resulting trimmed reads were aligned to the S. pombe reference genome (ASM294v2.27) and a previously generated GTF file using STAR (v2.7.10b) [81]. FeatureCounts from the Rsubread (v2.16.0) R package [82] was used to generate the raw count matrix which was then used as input for differential expression analysis with the edgeR (v3.42.4) R package [83]. Genes with a log2 fold change (Log2FC) greater than 1 or less than − 1, and a False Discovery Rate (FDR) less than 0.05, were considered differentially expressed.

ChIP and DRIP

ChIP and DRIP experiments were performed essentially as described previously [84]. Briefly, 2.5 × 10⁸ cells per IP were fixed in 1% formaldehyde for 15 min at room temperature. Cells were lysed using a bead beater (Biospec products, 2 x 2 min) and sonicated using a Bioruptor (Diagenode) for a total of 20 min (30 s on/30 s off on ‘high’ power). Immunoprecipitation was then performed overnight at 4°C, using 1 μl per IP of monoclonal anti-H3K9me2 (5.1.1 [85]) for ChIP, or 1 μl of monoclonal S9.6 anti-DNA-RNA hybrid (MABE1095, MerckMilipore) for DRIP. Immunoprecipitated DNA was recovered using Chelex-100 resin (BioRad), and quantified by qPCR using LightCycler 480 SYBR Green (Roche) and primers listed in S4 Table. Relative enrichments were calculated as the ratio of product of interest to control product (act1⁺) in IP over input. In all cases, histograms represent three biological replicates and error bars represent one S.D.

Immunoprecipitation

For IP-western, cultures were grown to mid-log phase in YES and 3 x 10⁸ cells were harvested, washed and resuspended in lysis buffer (50 mM Hepes pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.1% NP-40, 1 mM PMSF, 1× yeast protease inhibitor cocktail [Sigma]). Cells were lysed using a bead beater (Biospec products, 2 x 2 min), DTT was added to a final concentration of 500 μM, and the supernatant was recovered by 2 x 15 min centrifugation at 17,000 rcf at 4°C. Extracts were pre-cleared by incubation with pre-equilibrated protein G agarose for 1 hour at 4°C, and then incubated with pre-equilibrated protein G agarose plus anti-FLAG M2 (F3165, Merck, 1 μl/IP), or anti-GFP (A-11122, ThermoFisher, 5 ul/IP), for 3 hours at 4°C. The beads were washed three times with lysis buffer, and proteins eluted in 2x SDS sample buffer (50mM Tris-HCl pH 6.8, 2mM EDTA, 10% glycerol, 2% SDS, 2% β-Mercaptoethanol, 0.03% Bromophenol Blue) by boiling for 5 min. For analysis, proteins were separated by SDS-PAGE and transferred onto 0.45μm pore Protran nitrocellulose membrane (GE Healthcare) using semi-dry transfer apparatus (Hoeffer). Membranes were probed with mouse anti-FLAG M2 (F3165, Merck, 1:1000 dilution) or rabbit anti-GFP (A-11122, ThermoFisher, 1:1000 dilution). Secondary antibodies were IRDye conjugated anti-mouse or anti-rabbit (926-32210 or 926-68171, Li-Cor, 1:20,000 dilution) and imaging was performed using the Odyssey CLx Imaging System (LI-COR Biosciences).

Immunoaffinity purifications for mass spectrometry analysis were performed essentially as described previously [86]. Briefly, 500 ml cultures grown to a cell density of 10⁸ cells/ml in 4× concentrated YES media were milled in solid phase, and the cell powder solubilised in 10ml lysis buffer (50 mM HEPES–NaOH pH 7.5, 150 mM KCl, 0.1% NP-40, 0.2 mM PMSF, 0.2mM benzamidine, 1× EDTA-free protease inhibitor cocktail [Roche]). Following clarification by two rounds of centrifugation, immunoprecipitations performed by incubating with 8 μl Protein G Dynabeads (Novex) and 12 μl anti-FLAG antibody (F3165, Merck) per sample for 90 min at 4°C. The immunoprecipitated material was washed with lysis buffer, incubated in lysis buffer containing 2 mM MgCl₂ and 500 U of Benzonase nuclease (Novagen) for 15 min at 4°C, washed again with lysis buffer, and then eluted by incubation with 50 μl of 0.1% Rapigest (Waters) in 50 mM Tris-HCl pH 8.0 for 10 min at 50°C prior to LC-MS/MS analysis.

RNA-IPs were performed as above but with modified lysis buffer (50 mM HEPES–NaOH pH 7.5, 150 mM NaCl, 1mM MgCl₂, 0.1% NP-40, 5 mM DTT, 0.5 mM PMSF, 1× EDTA-free protease inhibitor cocktail [Roche], 0.2U/μl RNAsin Ribonuclease Inhibitor [Promega]), and excluding Benzonase treatment. After initial washes, immunoprecipitated material was resuspended in RNA extraction buffer (25 mM Tris–HCl pH 7.5, 5 mM EDTA pH 8, 50 mM NaCl, 0.5% SDS) with 200 ng/ml proteinase K and incubated at 37°C for 2 h. RNA was extracted with phenol:chloroform, precipitated with ethanol and 1 μl glycogen, and analysed by RT-qPCR.

Mass spectrometry

IP eluate was reduced with 25 mM DTT at 80°C for 1 min, then denatured by addition of urea to 8M. Sample was applied to a Vivacon 30k MWCO spin filter (Sartorius, VN01H21) and centrifuged at 12,500 g for 15 min. Protein retained on the column was then alkylated with 100 μl of 50 μM iodoacetamide (IAA) in buffer A (8 M urea,100 mM Tris pH 8.0) in the dark at RT for 20 min. The column was then centrifuged as before, and washed with 100 μl buffer A, then with 2 x 100 μl volumes of 50 mM ammonium bicarbonate (ABC). 3 μg/μl trypsin (Pierce, 90057) in 0.5 mM ABC was applied to the column and incubated at 37° for 16 hr. Digested peptides were then spun through the filter, acidified with trifluoroacetic acid (TFA) to pH <= 3, loaded onto manually-prepared and equilibrated C₁₈ reverse-phase resin stage tips (Sigma, 66883-U) [87], washed with 100 μl 0.1% TFA, and stored at -20°C.

For MS analysis, peptides were eluted in 40 μl of 80% acetonitrile in 0.1% TFA and concentrated down to 2 μl by vacuum centrifugation (Concentrator 5301, Eppendorf). The peptide sample was then prepared for LC-MS/MS analysis by diluting it to 5 μl by 0.1% TFA. LC-MS-analyses were performed on an Orbitrap Exploris 480 Mass Spectrometer (Thermo Scientific) coupled on-line, to an Ultimate 3000 RSLCnano Systems (Dionex, Thermo Fisher Scientific). Peptides were separated on a 50 cm EASY-Spray column (Thermo Scientific), which was assembled on an EASY-Spray source (Thermo Scientific) and operated at 50°C. Mobile phase A consisted of 0.1% formic acid in LC-MS grade water and mobile phase B consisted of 80% acetonitrile and 0.1% formic acid. Peptides were loaded onto the column at a flow rate of 0.3 μl per min and eluted at a flow rate of 0.25 μl per min according to the following gradient: 2 to 40% mobile phase B in 150 min, and then to 95% in 11 min. Mobile phase B was retained at 95% for 5 min and returned back to 2% a minute after until the end of the run (190 min). FTMS spectra were recorded at 120,000 resolution (scan range 350-1500 m/z) with an ion target of 3.0×10⁶ and maximum injection time of 50 ms. For the MS2 the resolution was set at 15,000 with ion target of 8.0×10⁴ and HCD fragmentation [88] with normalized collision energy of 30. The isolation window in the quadrupole was 1.4 Thomson. Only ions with charge between 2 and 6 were selected for MS2.

The MaxQuant software platform [89] version 1.6.1.0 was used to process the raw files and search was conducted against the Schizosaccharomyces pombe complete/reference proteome (Pombase, released in July 2017), using the Andromeda search engine [90]. For the first search, peptide tolerance was set to 20 ppm while for the main search peptide tolerance was set to 4.5 pm. Isotope mass tolerance was 2 ppm and maximum charge to 7. Digestion mode was set to specific with trypsin allowing a maximum of two missed cleavages. Carbamidomethylation of cysteine was set as fixed modification. Oxidation of methionine was set as variable modification. Absolute protein quantification was performed as described previously [91]. Peptide and protein identifications were filtered to 1% FDR.