Advertisement
  • Loading metrics

The intron in centromeric noncoding RNA facilitates RNAi-mediated formation of heterochromatin

  • Masatoshi Mutazono,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Misato Morita,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Chihiro Tsukahara,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Madoka Chinen,

    Current address: API Co., Ltd., Kanohsakurada-cho, Gifu, Japan.

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Shiori Nishioka,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Tatsuhiro Yumikake,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Kohei Dohke,

    Affiliation Laboratory of Chromosome Function and Regulation, Graduate School of Frontier Biosciences, Osaka University, Osaka, Japan

  • Misuzu Sakamoto,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Takashi Ideue,

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

  • Jun-ichi Nakayama,

    Affiliations Graduate School of Natural Sciences, Nagoya City University, Nagoya, Japan, Division of Chromatin Regulation, National Institute for Basic Biology, Okazaki, Japan

  • Kojiro Ishii,

    Affiliation Laboratory of Chromosome Function and Regulation, Graduate School of Frontier Biosciences, Osaka University, Osaka, Japan

  • Tokio Tani

    ttani@sci.kumamoto-u.ac.jp

    Affiliation Department of Biological Sciences, Graduate School of Science and Technology, Kumamoto University, Kurokami, Chuo-ku, Kumamoto, Japan

    ORCID http://orcid.org/0000-0003-4407-5541

The intron in centromeric noncoding RNA facilitates RNAi-mediated formation of heterochromatin

  • Masatoshi Mutazono, 
  • Misato Morita, 
  • Chihiro Tsukahara, 
  • Madoka Chinen, 
  • Shiori Nishioka, 
  • Tatsuhiro Yumikake, 
  • Kohei Dohke, 
  • Misuzu Sakamoto, 
  • Takashi Ideue, 
  • Jun-ichi Nakayama
PLOS
x

Abstract

In fission yeast, the formation of centromeric heterochromatin is induced through the RNA interference (RNAi)-mediated pathway. Some pre-mRNA splicing mutants (prp) exhibit defective formation of centromeric heterochromatin, suggesting that splicing factors play roles in the formation of heterochromatin, or alternatively that the defect is caused by impaired splicing of pre-mRNAs encoding RNAi factors. Herein, we demonstrate that the splicing factor spPrp16p is enriched at the centromere, and associates with Cid12p (a factor in the RNAi pathway) and the intron-containing dg ncRNA. Interestingly, removal of the dg intron, mutations of its splice sites, or replacement of the dg intron with an euchromatic intron significantly decreased H3K9 dimethylation. We also revealed that splicing of dg ncRNA is repressed in cells and its repression depends on the distance from the transcription start site to the intron. Inefficient splicing was also observed in other intron-containing centromeric ncRNAs, dh and antisense dg, and splicing of antisense dg ncRNA was repressed in the presence of the RNAi factors. Our results suggest that the introns retained in centromeric ncRNAs work as facilitators, co-operating with splicing factors assembled on the intron and serving as a platform for the recruitment of RNAi factors, in the formation of centromeric heterochromatin.

Author summary

Formation of centromeric heterochromatin is required for correct segregation of sister chromatids during mitosis. In fission yeast, formation of heterochromatin at centromeres is performed through the RNA interference (RNAi) system, which involves processing of noncoding RNAs transcribed from the centromeres. We found that the intron in the centromeric dg ncRNAs facilitates formation of centromeric heterochromatin in fission yeast. We showed that the splicing factor spPrp16p associates with the RNAi factor and intron-containing dg ncRNA. Removal of or mutations in the dg intron significantly decreased H3K9 dimethylation, suggesting that the intron and associated splicing factors serve as a platform for recruitment of RNAi factors. Inefficient splicing is a hallmark of intron-containing centromeric ncRNAs. Such repression of splicing seems to be important for facilitation of heterochromatin formation. Introns in euchromatic regions are removed by splicing to generate functional RNAs, whereas centromeric introns are retained in ncRNAs by splicing repression and play roles in gene silencing. Our findings shed light on the novel roles of introns in epigenetic regulation of gene expression and heterochromatin formation.

Introduction

Chromosome segregation is a fundamental process in the transmission of genetic information to the daughter cells in eukaryotic cells. During mitosis, formation of heterochromatin at centromeres is essential for correct segregation of chromosomes, because centromeric heterochromatin provides an environment that promotes the assembly of the kinetochore, the protein complex that serves as an attachment site for microtubules [1].

The RNA interference (RNAi) machinery in the fission yeast Schizosaccharomyces pombe was previously reported to be involved in the formation of centromeric heterochromatin [24]. Noncoding RNAs (ncRNAs) are transcribed from centromeres, which consist of repetitive sequences named dg and dh, by RNA polymerase II. Double-stranded RNAs (dsRNAs) are synthesized from transcribed ncRNAs by RNA-directed RNA polymerase (Rdp1p), and are then processed into small interfering RNAs (siRNAs) by Dicer. siRNAs then associate with Ago1p to form the RNAi-induced Transcriptional Silencing (RITS) complex with Chp1p and Tas3p [2], and function as guide molecules that direct the RITS complex to nascent ncRNAs through base pairing interactions. The RITS complex recruits the CLRC complex, which contains the methyltransferase Clr4p, to the pericentromere region, where it promotes the dimethylation of histone H3 lysine 9 (H3K9me2). Dimethylation of H3K9 creates a binding site for Swi6p, a heterochromatin protein 1 (HP1) homologue, to form heterochromatin.

Recent studies reported that mutations in specific splicing factors, such as Cwf10p (Complexed with Cdc5) and Prp10p (pre-mRNA processing 10), and spliceosomal U4 small nuclear RNA (snRNA) decreased the generation of siRNAs from centromeric ncRNAs, resulting in the defective formation of heterochromatin at centromeres [5, 6]. These findings suggest that the splicing machinery is involved in RNAi-mediated heterochromatic gene silencing at centromeres. Regarding the role of the splicing machinery in RNAi-directed heterochromatin formation, we found an mRNA-type intron in centromeric dg ncRNA and proposed a model in which the spliceosome (or sub-spliceosome consisting of a part of the splicing factors) that assembles on the intron in centromeric ncRNA serves as a platform for the binding of RNAi factors to centromeric ncRNAs, thereby promoting RNAi-mediated formation of centromeric heterochromatin [6]. On the other hand, it was recently reported that proper splicing of pre-mRNAs encoding RNAi factors is necessary for heterochromatin formation at the pericentromere region, suggesting that defective formation of centromeric heterochromatin in prp splicing mutants is a secondary effect of defective splicing of pre-mRNAs encoding RNAi factors [7].

In the present study, we showed that the splicing factor encoded by the prp16+ gene is enriched at the centromeres and interacts with Cid12p, a component of the RNA-dependent RNA polymerase complex (RDRC), in S. pombe. We also showed that the intron in dg centromeric ncRNA plays an important role in promoting the RNAi-mediated formation of centromeric heterochromatin. Our results indicate that the splicing machinery and intron in centromeric ncRNA play intrinsic roles in the formation of centromeric heterochromatin.

Results

prp14 exhibits defects in formation of centromeric heterochromatin

prp14 was isolated as a pre-mRNA splicing mutant that accumulates pre-mRNAs at non-permissive temperatures in the fission yeast Schizosaccharomyces pombe [8]. It exhibits cold sensitivity (22°C) and temperature sensitivity (37°C) for growth, and grows slowly at the permissive temperature of 33°C [8]. Complementation cloning of the prp14+ gene revealed that the mutation in prp14 resides in SPBC1711.17, which encodes an ATP-dependent RNA helicase homologous to Saccharomyces cerevisiae and human Prp16p (See S1 Text and S1 Fig), indicating that the prp14 mutation is a prp16 allele. Hereafter, we refer to prp14 and its wild-type product as prp16 and spPrp16p, respectively.

Some splicing mutants are defective in the RNAi-mediated formation of centromeric heterochromatin in S. pombe [5, 6], and we proposed a model in which the spliceosome or sub-spliceosome forms a platform for the recruitment of RNAi factors to ncRNAs transcribed from the centromere [6]. We found that prp16-2 was hypersensitive to the microtubule-destabilizing drug thiabendazole (TBZ) (Fig 1A) and accumulated unprocessed dg ncRNA transcribed from centromeres [6], suggesting that prp16 has defects in the attachment of microtubules to kinetochores and formation of centromeric heterochromatin.

thumbnail
Fig 1. prp16 has a defect in chromosome segregation.

(A) prp16-2 is hypersensitive to TBZ. Serially diluted cells indicated on the left of the panel were spotted on MMAU plates without TBZ (-TBZ) or with 10 μg/ml TBZ (+TBZ), and then incubated at 33°C for 5 days. Δago1 was spotted as a control; this strain has a defect in RNAi-induced formation of centromeric heterochromatin. (B) prp16-2 yields lagging chromosomes. Cells were double-stained with DAPI (red in the merged image) and antibody against tubulin (TAT1, green). Arrows indicate the lagging chromosome. The graph under the pictures shows the frequency of lagging chromosomes in late-anaphase cells in 972 (WT), prp16-2, or Δdcr1. Cells from three independent cultures at 33°C were analyzed. More than 50 late-anaphase cells were assessed for lagging chromosomes in each culture.

https://doi.org/10.1371/journal.pgen.1006606.g001

We further characterized prp16 in terms of the relationship between the splicing machinery and centromeric gene silencing. As shown in Fig 1B, the prp16-2 mutant exhibited a high incidence of lagging chromosomes during mitosis, consistent with defective attachment of microtubules to kinetochores. To examine the integrity of pericentromeric heterochromatin, which is required for the proper assembly of kinetochores, we constructed prp16-2 with the ura4+ marker gene in the otr1 region of centromere 1 (S2 Fig), and then serially spotted the resultant cells onto plates with or without 5-fluoroorotic acid (5-FOA). As shown in Fig 2A, prp16-2 cells harboring the inserted ura4+ gene were highly sensitive to 5-FOA, similar to the swi6 heterochromatin protein mutant, suggesting that, in these mutants, the ura4+ gene was expressed when inserted in centromeric regions. By contrast, a wild-type strain harboring the ura4+ gene inserted into the otr1 region grew well on 5-FOA plates. RT-PCR analysis also revealed that the ura4+ gene inserted into the otr1 region was expressed as in a swi6 mutant harboring the same insertion (Fig 2B).

thumbnail
Fig 2. prp16-2 exhibits defective formation of centromeric heterochromatin.

(A) Serially diluted cells harboring the ura4+ gene inserted in the otr1 region were spotted on plates containing 1 mg/ml 5-fluoroorotic acid (FOA). prp16-2 and swi6 mutants with the ura4+ gene were sensitive to 5’-FOA, indicating expression of the ura4+ marker gene inserted in the centromere region. The site of the ura4+ insertion is shown schematically in S2 Fig. (B) Expression of the ura4+ gene inserted in the centromere region in prp16-2 and swi6 was confirmed by RT-PCR. Ura4DS/E indicates a band derived from the authentic ura4+ gene, which harbors a 269 bp deletion. (C) ChIP analysis of H3K9me2 at the dg region was performed with the indicated strains using the anti-H3K9me2 antibody. Δdcr1 is a mutant deleting the dcr1+ gene essential for the RNAi pathway. (D) Northern blot analysis of siRNA derived from centromeric dg ncRNA revealed a reduction in the amount of centromeric siRNAs in prp16-2. SnoR69 RNA was used as a loading control. Lane 1: HM123 (wild-type); 2: Δdcr1; 3: HM123 transformed with pSP1; 4: HM123 transformed with pSP1-prp16+; 5: prp16-2 transformed with pSP1; 6: prp16-2 transformed with pSP1-prp16+.

https://doi.org/10.1371/journal.pgen.1006606.g002

To further demonstrate the impaired formation of centromeric heterochromatin in prp16, we performed a chromatin immunoprecipitation (ChIP) analysis using an anti-H3K9 dimethylation (H3K9me2) antibody. As shown in Fig 2C, the prp16 mutation decreased the H3K9me2 level in the pericentromeric region to a level observed in Δdcr1, indicating defective formation of pericentromeric heterochromatin in prp16. In addition, northern blot analysis revealed that the amount of centromeric siRNAs derived from dg ncRNA was markedly reduced in prp16-2 (Fig 2D, lane 5), consistent with our previous report showing that prp16 accumulates unprocessed dg ncRNA [6]. These results suggested that processing of centromeric ncRNAs is defective in prp16. Furthermore, expression of the wild-type prp16+ gene rescued defective production of centromeric siRNAs, indicating that the prp16+ gene is responsible for formation of centromeric heterochromatin (Fig 2D, lane 6).

spPrp16p is enriched at the centromere in S. pombe

Because prp16 was isolated in the screen for defective pre-mRNA splicing at the non-permissive temperature, the impaired formation of centromeric heterochromatin may have been caused by defects in splicing of pre-mRNAs encoding factors involved in RNAi-mediated formation of centromeric heterochromatin, as suggested by Kallgren et al. [7]. A RT-PCR analysis of splicing defects revealed that the prp16-2 mutation impaired splicing of ago1+, sir2+, hrr1+, arb2+, ers1+, and dsh1+ pre-mRNAs at the temperatures at which analyses on heterochromatin formation were conducted (S3 Fig). This result suggests that the observed defects in the formation of centromeric heterochromatin in prp16 are potentially due to splicing defects. On the other hand, prp13-1, which has defects in the formation of pericentromeric heterochromatin [6], similar to prp16, did not exhibit defects in pre-mRNA splicing for the RNAi factors analyzed, although very weak splicing defects were observed for Ago1p and Arb2p (S3 Fig, lanes prp13-1). In prp13-1, mature mRNAs encoding the RNAi factors were produced at levels almost equal to those in wild-type cells at all temperatures tested, suggesting that the defects observed in centromeric gene silencing in the prp mutants cannot be simply attributed to secondary effects of splicing defects.

Interestingly, spPrp16-GFP was detected in the nucleus as a dot-like signal with a diffuse nuclear distribution when it was over-expressed from a multicopy plasmid pSP1 (Fig 3A), but not when it was expressed from the endogenous locus (S4A Fig). Because prp16 exhibited a defect in formation of centromeric heterochromatin, we speculated that the nuclear site of the dot-like signal of overexpressed spPrp16-GFP corresponded to centromeres or kinetochores. To test this idea, we co-expressed spPrp16-GFP and Nuf2-RFP, a kinetochore protein tagged with RFP, and observed their localization. The dot-like signal of over-expressed spPrp16-GFP was colocalized with the Nuf2-RFP signal, suggesting that spPrp16p was enriched at the centromere (or kinetochore) when over-expressed (Fig 3B). In addition, ChIP analysis using a strain expressing FLAG-tagged spPrp16p from the endogenous locus revealed that spPrp16p was enriched at the dg and dh repeats in the pericentromere region (Fig 3C). These results suggest that spPrp16p plays some roles at the centromeres, in addition to its function as an RNA helicase in the splicing reaction [9].

thumbnail
Fig 3. spPrp16p associates with RNAi factors.

(A) spPrp16-GFP expressed from the multicopy plasmid exhibited a dot-like signal in the nucleus, in addition to diffuse nuclear distribution. Cells were cultured at 33°C and counterstained with Hoechst 33342. In the merged image, green and red indicate spPrp16p-GFP and Hoechst 33342 (DNA), respectively. (B) Enrichment of spPrp16p at the centromere region. Dot-like signals of spPrp16-GFP expressed from the plasmid and Nuf2-RFP, a kinetochore protein, expressed from its endogenous locus were colocalized in the nucleus, suggesting that spPrp16p is enriched at the kinetochore or centromere when overexpressed. Of 161 cells exhibiting co-expression of spPrp16-GFP and Nuf2-RFP, 42 cells (26%) showed dot-like signals of spPrp16-GFP and all of them colocalized with the Nuf2-RFP dot signals. Cells were stained with Hoechst 33342 before observations. (C) ChIP analysis of spPrp16-FLAG expressed from the endogenous locus at the dg and dh repeats. Enrichment of spPrp16-FLAG was calculated relative to the actin gene, which contains no intron. Cells with (+) or without (-) the plasmid expressing FLAG-tagged spPrp16p were used in these analyses. (D) RNA-independent interaction between spPrp16-HA and Cid12-TAP. A co-immunoprecipitation assay was performed using samples treated with (+) or without (-) RNase A. Cid12-TAP was immunoprecipitated using IgG–Sepharose beads (IP: IgG beads). Sepharose beads with no antibody were used as negative controls for immunoprecipitation. Extracts were prepared from cells expressing spPrp16p-HA and Cid12-TAP (spPrp16-HA: +) or only Cid12-TAP (spPrp16-HA: -). Western blot analysis was performed with an anti–protein A antibody (Cid12-TAP) or anti-HA antibody (spPrp16-HA).

https://doi.org/10.1371/journal.pgen.1006606.g003

spPrp16p interacts with Cid12p, a component of RDRC

To determine the role of spPrp16p in the formation of centromeric heterochromatin, we examined the physical interactions between spPrp16p and components of the RNAi pathway. To this end, we performed a co-immunoprecipitation assay using a strain expressing spPrp16p-HA and Cid12p-TAP or Chp1p-FLAG. Cid12p is a poly(A) polymerase family member protein that is a component of the RDRC, which binds centromeric ncRNAs to initiate dsRNA synthesis, and Chp1p is a component of the RITS complex that contains Ago1p [4, 10]. Although we were unable to detect an interaction between spPrp16p-HA and Chp1p-FLAG under the conditions tested (S4B Fig), we did detect spPrp16p-HA in the precipitate of Cid12-TAP, suggesting that spPrp16p interacts with Cid12p (Fig 3D, lanes RNase A-).

To determine whether the interaction between spPrp16p and Cid12p is RNA-dependent, we treated samples with RNase A and tested the interaction by co-immunoprecipitation. Treatment of samples with RNase A completely degraded dg RNAs (S4C Fig). As shown in Fig 3D (lanes marked as RNase A “+”), RNase treatment did not affect co-immunoprecipitation of Cid12p with spPrp16p, indicating that the association of spPrp16p with Cid12p is not mediated through RNA.

The intron in dg ncRNA facilitates the formation of heterochromatin

The physical association of spPrp16p with Cid12p supports the model proposed in our previous study, in which the spliceosome or sub-spliceosome assembled on the intron serves as a platform for recruiting RNAi factors [6]. To test the platform model further, we next investigated whether removal of the intron in dg ncRNA affects heterochromatin formation at the centromere using a minichromosome with or without the dg intron (Fig 4A). This minichromosome contains a central core sequence (cc2) and the part of the dg region necessary for centromere functions [11]. Interestingly, ChIP analyses of H3K9me2 and Swi6p heterochromatin protein using minichromosome-specific primers revealed that H3K9me2 and Swi6p levels on the dg region in the intron-less minichromosome (Less) were clearly lower than those in the intron-containing minichromosome (Full) (Fig 4B). In a strain harboring deletions of the dcr1+, clr4+, or rdp1+ genes, which are essential for the RNAi pathway, the level of H3K9me2 in the minichromosome was markedly reduced, demonstrating that dimethylation of H3K9 on the minichromosome is also mediated by the RNAi-mediated pathway. These results suggest that the dg intron facilitates the formation of heterochromatin.

thumbnail
Fig 4. Removal of the intron from the dg transcript impairs the formation of heterochromatin.

(A) Schematic representation of the minichromosomes. The intron-containing minichromosome (Full; pH-cc2) [11, 30] contains the cc2 (central domain) region, ura4+ gene, sup3-e gene, and a 2.1 kb KpnI–KpnI fragment derived from the otr dg region. The intron-less minichromosome (Less) lacks the dg intron. The arrow indicates the position of the minichromosome-specific reverse primer. (B) ChIP analyses of H3K9me2 (left panel) and Swi6p (right panel) at the dg sequences on the minichromosomes in the indicated strains. Quantitative PCR was performed using minichromosome-specific primers.

https://doi.org/10.1371/journal.pgen.1006606.g004

We also tested the association of spPrp16p with dg ncRNA with or without the intron (Fig 5A). RNA immunoprecipitation (RIP) analysis using minichromosome-specific primers revealed that the amount of intron-less dg ncRNA bound to spPrp16p-Myc was lower than that of intron-containing dg ncRNA (Fig 5A). Gel electrophoresis of RT-PCR products from the precipitates indicated that spPrp16p bound both unspliced and spliced dg ncRNAs (S5 Fig, WT/anti-Myc). The amount of steady-state intron-containing dg ncRNAs was lower than that of intron-less dg ncRNA (Fig 5B, lanes WT). There was no difference in the amount of intron-less dg ncRNA between the WT and Δdcr1, whereas the amount of intron-containing dg ncRNA was significantly elevated in Δdcr1 (Fig 5B, lanes Δdcr1). If the amounts of both transcripts were the same, these results imply that dg transcripts from the intron-containing minichromosome were more efficiently processed into siRNAs than those from the intron-less minichromosome. In addition, the decreased binding of spPrp16p to dg ncRNA was observed in Δdcr1 (Fig 5A), suggesting that the interaction between spPrp16p and dg ncRNA was dependent on a functional RNAi system.

thumbnail
Fig 5. RIP and mitotic stability analysis.

(A) Intron-less dg ncRNA bound spPrp16p less efficiently. RNA immunoprecipitation (RIP) analysis was carried out using the anti-Myc antibody and extracts prepared from wild-type or Δdcr1 cells possessing the intron-containing (Full) or intron-less (Less) minichromosome. spPrp16-Myc was expressed from the endogenous locus. qRT-PCR was performed using minichromosome-specific primers. Relative values were calculated as a ratio of IP/WCE. spPrp16p without the Myc-tag (untagged) was used as a negative control. (B) Amounts of steady-state dg ncRNA expressed from the minichromosome were analyzed by RT-PCR. Reverse transcription was performed using minichromosome-specific primers, and PCR was performed using primers that amplify the region spanning the intron (dg ncRNA-1) or region downstream of the intron (dg ncRNA-2). act1 mRNA was also detected as a loading control (act1 mRNA). “RT-” represents the reaction without reverse transcriptase. The positions of unspliced and spliced dg ncRNAs are indicated on the right of the panel. The amount of intron-less dg transcripts (Less) was larger than the amount of intron-containing dg transcripts (Full). In cells lacking the dcr1+ gene (Δdcr1), the amount of the intron-containing dg transcripts was elevated. (C) The mitotic stability assay of the minichromosome with (Full) or without (Less) the dg intron in wild-type or Δdcr1 cells. The p-value was calculated by a two-sample test for the equality of proportions with continuity corrections.

https://doi.org/10.1371/journal.pgen.1006606.g005

Previous studies reported that the formation of centromeric heterochromatin is important for preventing the missegregation of chromosomes during mitosis and meiosis [12, 13]. To evaluate the role of the dg intron in the formation of centromeric heterochromatin, we performed mitotic stability assays on minichromosomes. Specifically, we measured the frequency of mitotic losses in the minichromosome with or without the dg intron. Because the minichromosome used in these experiments contained the leu1 selection marker, mitotic loss of the minichromosome could be analyzed by growing minichromosome-containing cells on plates with or without leucine. The results obtained indicated that the frequency of missegregation of the intron-less minichromosome (Less) was higher than that of the intron-containing minichromosome (Full) in wild-type cells (Fig 5C). We also observed the same phenomenon in Δdcr1 cells, suggesting that the effect of the dg intron on chromosome segregation is at least partially RNAi-independent. Collectively, these results led us to conclude that the intron in dg ncRNA facilitates formation of centromeric heterochromatin in S. pombe.

Analysis of a chimeric RNA containing the intron of the euchromatic gene

To examine the role of the dg intron in facilitating H3K9 dimethylation, we replaced the dg intron in the dg ncRNA gene (SPNCRNA.232) cloned in pREP1 (the 5’-long construct) with the intron from the gcd10 gene, which encodes a tRNA methyltransferase (Fig 6A). We selected the gcd10 intron because the intron length (136 bp) is similar to that of the dg intron (138 bp). As shown in S6 Fig, the chimeric transcript from the fusion construct, 5’-long (g10in), was spliced with low efficiency, although the gcd10 intron itself is spliced efficiently in the original gcd10 pre-mRNA. ChIP analysis using plasmid-specific primers revealed that a H3K9me2 level on the 5’-long construct was significantly reduced in Δdcr1, Δclr4, and prp16-2, suggesting that enhanced methylation of H3K9 on the plasmid construct is RNAi- and spPrp16p-dependent, like the minichromosome constructs (Fig 6B). Interestingly, the 5’-long (g10in) construct also exhibited a reduced level of H3K9me2 relative to that of the 5’-long construct [Fig 6B, WT/5’-long and WT/5’-long (g10in)]. By contrast, histone H3 on the 5’-long (g10in) construct was almost the same as that on the 5’-long construct (Fig 6C). These results suggest that the sequence of the dg intron itself is important for H3K9 dimethylation.

thumbnail
Fig 6. The intron sequence is important for facilitated H3K9 dimethylation.

(A) Schematic structures of the 5’-long construct that expresses SPNCRNA.232 with the 570 bp upstream sequence, dg intron, and 106 bp downstream sequence and the 5’-long (g10in) construct containing the gcd10 intron replaced with the dg intron. The arrow indicates the position of the plasmid-specific reverse primer. (B) ChIP analysis of H3K9me2 on the plasmids. The level of H3K9me2 in the 5’-long construct was reduced in the RNAi mutants (Δdcr1 and Δclr4) and prp16-2. The 5’-long (g10in) construct had a reduced level of H3K9me2 in wild-type cells. (C) ChIP analysis of histone H3 on the 5’-long and 5’-long (g10in) constructs. qRT-PCR was carried out using plasmid-specific primers.

https://doi.org/10.1371/journal.pgen.1006606.g006

To identify the sequence element involved in the facilitated H3K9 dimethylation, we replaced a part of the dg intron in the 5’-long construct with the sequence of the gcd10 intron, as shown in S7A Fig, and performed ChIP analysis using the plasmid-specific primers. Replacement of the 5’ side sequence (46 bp) of the dg intron with that of the gcd10 intron decreased the level of H3K9me2, suggesting that an element responsible for facilitating H3K9 dimethylation is present in the 5’ region of the intron (S7B Fig).

Mutations at splice sites affect the efficiency of H3K9 methylation

To determine whether recognition of splice sites in the dg intron is required to facilitate heterochromatin formation, we constructed plasmids containing the dg intron with mutated splice sites. As shown in Fig 7A, we mutated the 5’ splice site (5’SS), branch and 3’ splice sites (BP+3’SS), or all of them (5’SS+BP+3’SS) in the dg intron. All of these mutations completely inhibited splicing of the dg transcripts (S8A Fig). Interestingly, ChIP analysis revealed that the level of H3K9me2 at the dg region was significantly reduced in plasmids containing mutations at the splice and branch sites, in comparison with plasmid containing the wild-type dg intron (Fig 7B). In addition, RIP analysis using a strain expressing Myc-tagged spPrp16p revealed that association of spPrp16p with the dg ncRNA was significantly decreased by the mutations at the 5’SS, BP+3’SS and 5’SS+BP+3’SS (S8B Fig). These results suggest that the splice and branch site sequences recognizable by splicing factors are necessary for the facilitation of H3K9 dimethylation, in addition to the 5’ cis-element in the dg intron.

thumbnail
Fig 7. Mutations at the splice and branch sites decrease the efficiency of H3K9 dimethylation.

(A) Schematic representation of mutation sites in the 5’-long construct. Mutated nucleotide sequences of the 5’ splice (5’SS), branch (BP), and 3’ splice (3’SS) sites are shown below the schematic. (B) ChIP analysis of H3K9me2 at the plasmids was performed using strains harboring plasmids with the indicated mutations. qPCR was performed using specific primers.

https://doi.org/10.1371/journal.pgen.1006606.g007

An upstream region is involved in the splicing repression of dg ncRNA

In S. pombe, more than 43% of genes contain introns [14]. However, in contrast to the dg intron, the introns in these euchromatic genes do not induce the formation of heterochromatin. What is the difference between the intron in the dg ncRNA gene and those in euchromatic genes? In this regard, we noted that the splicing efficiency of dg ncRNA is very low (S5 Fig, lane: Input; S6 Fig, lane: 5’-long; [6]), despite the fact that the splice and branch site sequences in the dg intron closely match the corresponding consensus sequences in S. pombe [15]. This low splicing efficiency appears to be a unique feature of the dg intron, as most introns in the euchromatic regions are efficiently spliced after transcription. Interestingly, we found that truncated dg ncRNA transcribed from a plasmid with short upstream and short downstream sequences driven by the nmt1 promoter was efficiently spliced, similar to euchromatic introns (S9A and S9B Fig, dg-short). This efficient splicing was not due to the nmt1 promoter, because original-size dg ncRNA transcribed under the control of the nmt1 promoter exhibited very low splicing efficiency, similar to endogenous dg ncRNA (S9A and S9B Fig, dg long). The plasmid producing efficiently spliced dg ncRNA exhibited a markedly reduced level of H3K9me2 in the ChIP analysis (S9C Fig, Short) relative to the plasmid possessing the long construct with low splicing efficiency (S9C Fig, Long), suggesting that retention of the dg intron by splicing repression facilitates dimethylation of H3K9.

Because a 5’-long construct with the long upstream sequence (-570 bp from the dg intron) and shortened downstream sequence (+106 bp from the dg intron) exhibited low splicing efficiency, similar to the wild-type dg construct (dg-long), we speculated that the upstream region contains the cis-element needed for splicing repression to retain the intron element facilitating H3K9me2. Because splicing repression of the 5’-long construct was not relieved in Δdcr1 and Δclr4, inhibition of dg RNA splicing is not dependent on the RNAi system (S10 Fig). To identify the cis-elements responsible for splicing repression, we constructed a series of plasmids expressing dg transcripts with serially deleted upstream regions (Fig 8A). The transformants with these plasmids were then subjected to RT-PCR, qRT-PCR and ChIP analyses. As a result, we found that deletion of the upstream region gradually increased splicing efficiency (Fig 8B) and decreased the H3K9me2 levels (Fig 8C), implying a tendency toward an inverse relationship between the splicing efficiency of the dg intron and H3K9 modification.

thumbnail
Fig 8. Deletion of the upstream region affects pre-mRNA splicing and H3K9 dimethylation.

(A) Schematic representation of dg fragments containing serial deletions in the upstream region. Constructs were fused to the nmt1 promoter and terminator (gray arrows and boxes, respectively). Numbers on the left represent the lengths of intron-containing dg DNA fragments and the Act1/dg chimeric fragment ligated to the nmt1 promoter and terminator. The 5’-long construct consists of the upstream region (570 bp), dg intron (138 bp), and downstream region (106 bp). The Act1/dg construct consists of the actin gene fragment (442 bp) and the dg-short fragment (390 bp). (B) Splicing of dg ncRNAs transcribed from serially deleted fragments was examined by RT-PCR (left panel) and qRT-PCR (right panel) using plasmid-specific primers. Products specific for spliced dg ncRNA were amplified using spliced product-specific primers. The ratios of the spliced products to the total transcripts (spliced and unspliced products) were calculated and graphed. Act1 mRNA was analyzed as a loading control. No bands were detected in samples without the reverse transcription reaction. (C) ChIP analysis of H3K9me2 at the dg region was performed with the H3K9me2 antibody and strains possessing the indicated plasmids. qPCR was performed using plasmid-specific primers. (D) Splicing of the Act1/dg chimeric transcript was analyzed by RT-PCR (upper panel) and qRT-PCR (lower panel). Splicing of 5’-long and dg-short transcripts were also examined as controls.

https://doi.org/10.1371/journal.pgen.1006606.g008

We also constructed a deletion series using the 5’-long (g10in) construct with the gcd10 intron and examined their splicing efficiencies. As shown in S11 Fig, deletions of the upstream region promoted splicing efficiency of the gcd10 chimeric constructs, although splicing enhancements in dg-4 and short chimeric transcripts were weak for unknown reasons. This result suggests that splicing repression is independent of intron sequences.

The gradual increase of splicing efficiency in the 5’-long deletion constructs suggests that the upstream region contains no specific cis-elements involved in repression, and that the distances between the transcription start sites and the intron are important for the splicing repression. To investigate this possibility, we ligated a 442 bp DNA fragment of the actin gene (+4 to +445) upstream of dg-short to yield the Act1/dg construct, which has almost the same upstream length as 5’-long but with a different sequence (Fig 8A). RT-PCR analysis revealed that the splicing efficiency of Act1/dg transcripts was very low, like the 5’-long transcripts (Fig 8D), indicating that distance from the transcription start site to the dg intron is important for splicing repression. Moreover, ChIP analysis revealed that the level of H3K9 dimethylation on the Act1/dg construct was low (Fig 8D), suggesting that elements involved in promoting formation of H3K9me2 are present not only in the dg intron but also in the 5' upstream region.

Intron-containing centromeric ncRNAs are inefficiently spliced

Recently, Lee et al. showed that “cryptic introns” specify heterochromatin domains (HOODs) in the S. pombe genome [16]. In particular, they reported the presence of the “cryptic introns” in the dh transcript and dg antisense transcript [16] (Fig 9A). We examined splicing of these intron-containing centromeric transcripts by RT-PCR and found that their splicing efficiencies were also low, like the dg intron, in wild-type cells (Fig 9B, WT). Interestingly, in a strain harboring deletion of the cid12+, dcr1+, or clr4+ genes, which are essential for the RNAi machinery, splicing of the antisense dg intron was significantly upregulated (Fig 9B, Δcid12, Δdcr1, and Δclr4), suggesting that its splicing is also repressed in wild-type cells, like the dg intron, and that the RNAi system is involved in that splicing repression. Overexpression of Dcr1p and Cid12p, by contrast, did not affect splicing of centromeric ncRNAs in wild-type cells (S12 Fig).

thumbnail
Fig 9. Inefficient splicing of centromeric introns.

(A) Schematic representation of introns in centromeric ncRNAs. Arrows indicate the direction of transcription. Vertical boxes denote the intron. (B) RT-PCR analyses of the dg and dh transcripts and dg antisense transcript containing the intron. Splicing efficiencies of the transcripts are low in wild-type cells. Deletion of the cid12+, dcr1+, and clr4+ genes, which are essential for RNAi-mediated formation of heterochromatin, promoted splicing of the dg antisense transcript. Act1 mRNA was detected as a loading control (act1 mRNA). “RT-” represents the reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.g009

Discussion

In this study, we cloned the S. pombe gene responsible for the prp14 mutation, which impairs RNAi-mediated formation of centromeric heterochromatin (Figs 1 and 2), as well as pre-mRNA splicing [8]. We found that the prp14+ gene encodes a homologue of the S. cerevisiae and human splicing factor Prp16p, a DEAH-box RNA helicase with ATPase activity [17] (S1C Fig). We, therefore, renamed prp14 as prp16. Although prp16 exhibited defective splicing of pre-mRNAs encoding several factors involved in RNAi-mediated gene silencing (S3 Fig), the observation that spPrp16p (encoded by the prp16+ gene) is enriched at the centromere and interacts with Cid12p, a component of the RDRC of the RNAi system, suggested that spPrp16p plays a direct role in the formation of centromeric heterochromatin (Fig 3). Thus, defective formation of centromeric heterochromatin in prp16 might be caused by composite effects of the splicing impairments and the functional abnormality of spPrp16p in heterochromatin formation.

We previously reported that another temperature-sensitive splicing mutant, prp13-1, in which the affected gene encodes the U4 snRNA essential for pre-mRNA splicing [18], is defective in centromeric gene silencing [6]. In the case of prp13, splicing impairments for RNAi factors were not observed at the temperature (26°C) that induced defective formation of centromeric heterochromatin (S3 Fig). Therefore, the phenotype defective in the formation of centromeric heterochromatin in prp13-1 does not appear to be a secondary effect of splicing defects, although we cannot exclude the possibility that pre-mRNA splicing of unknown factors involved in gene silencing is defective in prp13-1.

Several studies have reported crosstalk between splicing factors and RNAi- or exosome-mediated gene silencing. The first such evidence came from pull-down experiments using TAP-tagged Cid12p, which showed that several spliceosomal proteins co-precipitated with Cid12p-TAP in Δrdp1 [10]. Bayne et al. [5] revealed that csp4 (centromere: suppressor of position effect 4) and csp5, which alleviated the silencing of marker genes inserted into the otr region in the centromere, encode the splicing factors Cwf10p and Prp39p, respectively. They also demonstrated that Cwf10p associates with Cid12p and centromeric chromatin, suggesting that splicing factors facilitate RNAi-directed silencing [5]. Another splicing factor, Spf30p, was also shown to assist exosome-mediated centromeric gene silencing [19]. In addition, in the yeast Cryptococcus neoformans, spliceosomes stalled on intron-containing pre-mRNAs serve as a signal that induces siRNA synthesis by the SCANR (Spliceosome-Coupled And Nuclear RNAi) complex [20]. It remains unclear why these splicing factors associate closely with RNAi- or exosome-mediated gene silencing. In our previous study on prp13-1, we identified an intron typical for those in pre-mRNAs in centromeric dg ncRNA, and proposed a model in which the spliceosome or sub-spliceosome assembled on the intron functions as a platform for recruiting RDRC to facilitate the processing of centromeric ncRNAs [6]. In this study, we tested this platform model by removing the intron from the dg transcript and mutating the splice sites of the dg intron, leading to reduced binding of spPrp16p to dg ncRNA (Fig 5A and S8B Fig). The results showed that the removal of and mutations in the intron significantly decreased levels of H3K9me2 and Swi6p binding (Figs 4B and 7B), indicating that the intron is actually necessary for facilitating RNAi-mediated formation of heterochromatin.

Lee et al. recently suggested that the splicing machinery and its associated factor Nrl1p (NRDE-2 like 1) act on introns to specify domains for heterochromatin (HOODs) in S. pombe [16]. They discovered widespread previously unannotated “cryptic introns” in genes for mRNAs and ncRNAs, including the dh ncRNA gene in the pericentromere [16]. The dh ncRNAs, as well as dg ncRNAs, are processed to siRNAs essential for RNAi-mediated heterochromatin formation. They proposed that RNAi targets include “cryptic introns”, which play an important role in defining the targets of RNAi/exosome-mediated heterochromatin assembly through the spliceosome [16]. Our results and the aforementioned findings support the idea that centromeric introns serve as signals for heterochromatin formation, in collaboration with the splicing machinery.

We found that splicing of the dg ncRNA was affected by the distance from the transcription start site to the intron (Fig 8). The chimeric 5’-long (g10in) and Act1/dg constructs also exhibited low splicing efficiencies, supporting the idea that splicing repression of the dg intron depends on the length, rather than the nucleotide sequence, of the upstream sequence (S6 and S11 Figs and Fig 8D). Several lines of experiments in yeast and mammals demonstrated that a cap-binding complex (CBC) promotes binding of U1 snRNP to cap-proximal 5' splice sites and stimulates splicing; consequently, introns positioned closely to transcription start sites are spliced more efficiently [21, 22]. Consistent with this, in single-intron genes in yeasts, most introns are found at the 5' ends [23]. Thus, the downstream positioning of the dg intron may keep splicing efficiency low so that the intron is retained within the transcript.

qRT-PCR and ChIP analyses of the 5’ deletion constructs showed that splicing efficiency of the dg intron and H3K9me2 levels tended to have an inverse relationship (Fig 8B and 8C). Deletion of the upstream region increased splicing efficiency gradually; however, the H3K9me2 levels decreased abruptly after deletion of 73 bp from the 5’long construct, suggesting that there exists a threshold splicing efficiency for the enhancement of heterochromatin formation or that a cis-element facilitating heterochromatin formation, which functions in combination with the dg intron, is present in the 5’ upstream region.

We found that the “cryptic introns” also had very low splicing efficiencies (Fig 9). It is noteworthy that depletion of a factor essential for the RNAi machinery, such as Cid12p, Dcr1p, or Clr4p, facilitated splicing of the antisense dg transcript (Fig 9B). This result suggested that the active RNAi machinery is closely involved in splicing repression for the antisense dg intron. Interestingly, the antisense dg intron is localized within the hairpin-like, highly structured region called RevCen (Fig 10A) [24]. Small centromeric RNAs derived from the RevCen region were found in Δrdp1 cells, suggesting that hairpin-like RevCen transcripts are processed into siRNAs by Dcr1p, independently of RDRC activity [24]. These siRNAs have been proposed to function in initiation of heterochromatin formation [24]. It is possible that removal of the antisense dg intron from RevCen transcripts by splicing disrupts the hairpin-like structure necessary for RDRC-independent processing, resulting in reduced formation of heterochromatin.

thumbnail
Fig 10. Model showing how the intron in centromeric ncRNAs facilitates RNAi-mediated formation of heterochromatin.

(A) Predicted hairpin-like structures of RevCen [24] containing the antisense dg intron. Secondary structure of RevCen predicted by Mfold [25] was assessed experimentally [24]. Intron regions are indicated in red. (B) Efficient recruitment of RNAi components onto centromeric ncRNA is achieved through their interaction with the spliceosome (or sub-spliceosome) assembled on the intron. Repression of the splicing reaction is induced by the long upstream region (i) or induced by the active RNAi machinery, resulting in formation of a hairpin-like structure in the intron region, which might be processed into siRNAs RDRC-independently as proposed in [24]. Then, the stalled splicing complex may promote the association of RNAi components with ncRNAs, facilitating heterochromatin formation.

https://doi.org/10.1371/journal.pgen.1006606.g010

Furthermore, regions containing the dg and dh introns were also predicted by Mfold, which predicts RNA folding by minimizing free energy [25], to form highly structured hairpins (S13 and S14 Figs), although the predicted secondary structures should be confirmed experimentally, as was done previously for RevCen structure using enzymatic and chemical probing [24]. The centromeric introns retained in the transcripts by splicing repression may function not only in the splicing machinery-dependent facilitation of heterochromatin formation via assembly of the platform complex, but also in RDRC-independent initiation of heterochromatin assembly through formation of hairpin-like RNAs that are processed directly by Dcr1p as proposed for RevCen [24].

Splicing repression may result in formation of a stalled spliceosome, triggering association and recruitment of RNAi factors to centromeric ncRNAs and promoting the formation of heterochromatin (Fig 10B). Given that the 5’-long (g10in) and Act1/dg transcripts did not induce facilitated H3K9 dimethylation, the cis-elements in the dg intron and 5' upstream region seem to be required for upregulation of H3K9 dimethylation, in addition to repression of splicing (S7 Fig and Fig 8). Further analyses of the cis-elements and the mechanisms that connect splicing repression to formation of heterochromatin are now underway in our laboratory.

Interestingly, the human homologue of spPrp16p, DHX38, is a component of the Interphase Centromere Complex (ICEN) [26], suggesting that the human homologue of spPrp16p also plays an important role at the centromere. We recently found that satellite I ncRNAs transcribed from human centromeres associate with Aurora B kinase, a major regulator of chromosome segregation, and that knockdown of centromeric satellite I ncRNAs induces the defective attachment of microtubules to kinetochores, leading to impaired segregation of chromosomes in HeLa cells [27]. Association of DHX38 with satellite I ncRNAs is now under investigation. Analyses of spPrp16p, together with its human homologue DHX38, and introns in centromeric ncRNAs will shed light on evolutionarily conserved crosstalk between the splicing machinery and the mechanisms that maintain centromere functions through ncRNAs.

Materials and methods

S. pombe strains, plasmids, and general methods

The S. pombe strains and plasmids used in this study are listed in S1 and S2 Tables, respectively. The general methods used for analyses of S. pombe are described in refs. [28] and [29].

TBZ sensitivity test and silencing assay

For the TBZ sensitivity test, cells were suspended in YE broth at a density of 2 × 103 cells/ml, serially 5-fold diluted, and spotted on plates containing TBZ (10 μg/ml). In the silencing assay, serially diluted cells were spotted on N/S plates (non-selective YE medium supplemented with adenine, leucine, and uracil) or 5-FOA plates (N/S plates containing 1 mg/ml 5-FOA). The plates were incubated at 33°C for 3–4 days.

RT-PCR analysis

Cells were cultured to mid-log phase in YE or MMAU medium at 26°C [for 972(WT), Δdcr1, Δclr4, and Δcid12] or 33°C (for prp16-2). For the splicing analysis in S3 Fig, the indicated strains were cultured at 26°C (prp16-2, 30°C), and then shifted to 22°C or 30°C (prp16-2, 22°C or 26°C) for 6 hours. To isolate total RNA, cells were harvested by centrifugation and treated with phenol: chloroform (5:1). After treatment with DNase I (Ambion), reverse transcription (RT) was performed using 1 μg of total RNA and a cDNA synthesis kit (TaKaRa). Random primers were used for reactions with act mRNA and the mRNAs in S3 Fig. Specific primers were used for RT of dg and dh ncRNAs, transcripts from the minichromosomes, and plasmid constructs. RT samples were then subjected to PCR amplification using the specific primers listed in S3 Table.

Splicing efficiencies of dg ncRNAs with upstream deletions were quantitated by qRT-PCR of total and spliced dg ncRNAs. Primers corresponding to the dg second exon and downstream vector region (the nmt1 terminator region) were used for amplification of total products, and a primer spanning the exon junction and a primer complementary to the downstream vector region were used for amplification of the spliced products. The 5’-long plasmid without the intron was used for the qPCR standard for total and spliced products. The ratio of spliced dg ncRNA to total dg ncRNA was calculated and graphed. The assay was performed three times independently.

Chromatin Immunoprecipitation (ChIP) analysis

ChIP analysis of H3K9me2, Swi6p, and spPrp16p-Flag was performed as described [6]. Extracts were prepared from cells cultured at 33°C (prp16-2) or 26°C (Δdcr1, Δclr4, Δago1, and wild-type 972). Antibodies against H3K9me2 (ab1220) and FLAG (F3165) were purchased from Abcam and Sigma, respectively. Anti-Swi6p antibody was produced by Dr. Jun-ichi Nakayama. Real-time quantitative PCR was carried out with a Roche Diagnostic LightCycler using the primers listed in S3 Table. Data analyses were performed using the LightCycler Software (ver. 3.5). ChIP analyses were conducted at least three times independently.

RNA Immunoprecipitation (RIP) analysis

A minichromosome was introduced into cells expressing spPrp16-Myc from the endogenous locus. Cells (972 or Δdcr1) transformed with the indicated constructs were grown to mid-log phase in MMAU medium at 26°C, and then fixed with 1% formaldehyde for 30 min at room temperature. After treatment with 125 mM glycine, cells were washed with PBS and resuspended in lysis buffer (50 mM HEPES-KOH, pH 7.5, 140 mM NaCl, 1% TritonX-100, 0.1% DOC, 1 mM PMSF, complete EDTA-free protease inhibitor cocktail, and 100 U/ml RNase inhibitor). Cells were then disrupted using a Multi-beads shocker (Yasui Kikai), followed by sonication with a Bioruptor (BM Equipment). After treatment with DNase (Ambion), spPrp16-Myc was immunoprecipitated using anti c-Myc antibody (9E10, Santa Cruz), and RNAs were extracted from the precipitates. RT of RNAs was performed using the plasmid-specific primers listed in S3 Table. Real-time quantitative PCR and data analyses were performed as described above in the “ChIP analysis” section. RIP analyses were conducted at least three times independently.

Immunostaining

Immunostaining of tubulin in the prp16-2 mutant was performed as described previously [6]. The TAT1 monoclonal antibody against tubulin was provided by Dr. Keith Gull. After counterstaining with DAPI, cells were observed with an Olympus AX70 fluorescence microscope equipped with a Photometrics Quantix-cooled CCD camera.

Mitotic stability assay

Cells containing the minichromosome with or without the dg intron were cultured to mid-log phase in MMAU medium at 26°C. Cells (300 cells/plate) were plated on MMAU and MMALU plates, and then incubated at 26°C for 10 days. Colonies grown on the plates were counted to calculate the mitotic stability of the minichromosome. The number of colonies on MMAU plates was divided by the number of colonies on MMALU plates.

Supporting information

S1 Fig. Cloning of the prp14+ gene.

(A) Schematic structure of the genomic fragment in pSP#1. The subcloned regions and results of functional complementation are shown below the gene structure. The subcloned fragments that rescued defective growth at the non-permissive temperature are indicated as Ο, and non-rescuing fragments are indicated as X. (B) Complementation test of the subcloned fragments. The fragment containing the spPrp16p gene rescued the cold-sensitive phenotype. Prp16-pSP1 indicates the prp16 mutant transformed with pSP1 vector. (C) Amino acid alignment of S. pombe spPrp16p, S. cerevisiae Prp16p, and human DHX38/Prp16p. DEAH-box helicase and helicase superfamily C-terminal domains are underlined. Identical and similar amino acids conserved among species are colored in red. Amino acids conserved between two species are shown in blue.

https://doi.org/10.1371/journal.pgen.1006606.s002

(PDF)

S2 Fig. A schematic of the centromere region with the site of the ura4+ insertion used for the silencing assay.

The ura4+ gene was inserted into the otr1R region of chromosome 1.

https://doi.org/10.1371/journal.pgen.1006606.s003

(PDF)

S3 Fig. Analysis of defective splicing of pre-mRNAs encoding factors involved in RNAi-induced formation of centromeric heterochromatin.

Total RNAs were isolated from the indicated strains cultured at 22°C, 26°C, or 30°C and then subjected to RT-PCR analysis using primers specific for the indicated intron regions.

https://doi.org/10.1371/journal.pgen.1006606.s004

(PDF)

S4 Fig. Localization of spPrp16p-GFP expressed from the endogenous locus, co-immunoprecipitation and RNase treatment.

(A) spPrp16p-GFP expressed from the endogenous locus localized diffusely in the nucleus. No dot-like signals were observed. Cells cultured at 30°C were counterstained with Hoechst 33342 and observed through a Nikon Eclipse Ti fluorescence microscope equipped with an ORCA-R2 cooled CCD camera. (B) spPrp16-HA does not associate with Chp1p-FLAG. Co-immunoprecipitation analysis was carried out using a strain expressing spPrp16-HA and Chp1p-FLAG. spPrp16-HA was immunoprecipitated with (IP: anti-HA) or without (IP: -) an anti-HA antibody. The precipitates were then subjected to western blot analysis using anti-FLAG antibody. No band was detected in the precipitates with the anti-HA antibody, suggesting no association between spPrp16p and Chp1p. Arrowheads indicate positions of spPrp16-HA and Chp1p-FLAG. (C) Treatment of extracts with RNase A degraded dg ncRNA. Extracts from cells with (spPrp16-HA: +) or without (spPrp16-HA: -) a plasmid expressing spPrp16-HA were treated with 10 μg/ml RNase A at 37°C for 10 min (RNase A: +), and then subjected to the RT-PCR analysis. “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s005

(PDF)

S5 Fig. spPrp16p binds unspliced and spliced dg ncRNA.

RT-PCR products prepared from the indicated precipitates were electrophoresed on an 8% polyacrylamide gel and stained with ethidium bromide. RNA immunoprecipitation was performed using anti-Myc antibody. “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s006

(PDF)

S6 Fig. Splicing of the transcripts from the 5’-long and 5’-long (g10in), which contains the gcd10 intron, is not efficient.

RT-PCR was performed using plasmid-specific primers. Act1 mRNA was detected as a loading control (act1 mRNA). “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s007

(PDF)

S7 Fig. The 5’ region of the dg intron is important for facilitation of H3K9 dimethylation.

(A) Schematic representation of the chimeric 5’-long constructs, in which part of the intron was replaced with the sequence of the gcd10 intron (white boxes). (B) ChIP analysis of H3K9me2 on the chimeric 5’-long constructs. The 5’-long (g10in-A) construct in which the 5’ region of the dg intron was replaced with the corresponding region of the gcd10 intron had a reduced level of H3K9me2.

https://doi.org/10.1371/journal.pgen.1006606.s008

(PDF)

S8 Fig. Analysis of the dg ncRNAs with mutated splice sites.

(A) Splicing of the dg ncRNA was completely inhibited by the mutations introduced. RT was performed using plasmid-specific primers. PCR was conducted using primers that amplified the region spanning the intron (upper panel), which were used for evaluation of the splicing reaction, or primers that amplified the region downstream of the intron (middle panel), which were used to determine the amounts of the transcripts. act1 mRNA was analyzed as a loading control (bottom panel). “RT-” indicates a reaction without reverse transcriptase. (B) RIP analysis of spPrp16-Myc was performed with the anti-Myc antibody (anti-Myc) or non-immune IgG (normal IgG) and strains possessing plasmids with the indicated mutations. qPCR was performed using plasmid-specific primers.

https://doi.org/10.1371/journal.pgen.1006606.s009

(PDF)

S9 Fig. A dg transcript with the short flanking sequences exhibited more efficient splicing.

(A) Schematic representation of plasmids expressing dg transcripts. The dg-long and dg-short constructs contain the 1,230 and 390 bp fragments including the dg intron, in addition to the nmt1 promoter and terminator (gray arrow and box, respectively). (B) Splicing of dg ncRNAs expressed from plasmids or the endogenous locus was analyzed by RT-PCR using plasmid-specific primers (dg-long or dg-short) or endogenous gene-specific primers (endo. dg). The arrowheads indicate unspliced and spliced products. The dg-short transcript was spliced efficiently. No bands were detected in any samples without the reverse transcription reaction. act1 was detected as a loading control. (C) ChIP analysis of H3K9me2 at the dg region was performed using the H3K9me2 antibody and strains harboring the dg-long or dg-short plasmid. qPCR was performed using plasmid-specific primers.

https://doi.org/10.1371/journal.pgen.1006606.s010

(PDF)

S10 Fig. Inefficient splicing of the dg intron in the 5’-long construct in the RNAi-defective deletion mutants.

Splicing of the 5’-long dg intron in the indicated strains was analyzed by RT-PCR. Act1 mRNA was detected as a loading control (act1 mRNA). “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s011

(PDF)

S11 Fig. Deletion of the upstream region affects the splicing efficiency of the gcd10 chimeric transcript.

(A) Schematic representation of chimeric fragments containing the gcd10 intron and serially deleted upstream region. Constructs were fused to the nmt1 promoter and terminator (gray arrows and boxes). The numbers on the left denote the length of each intron-containing dg DNA fragment inserted between the nmt1 promoter and terminator. (B) Splicing of the gcd10 chimeric RNAs transcribed from serially deleted fragments was examined by RT-PCR (upper panel) and qRT-PCR (lower panel) using plasmid-specific primers. To draw a graph of splicing efficiency, the ratios of the spliced products to the total transcripts were calculated. In RT-PCR, Act1 mRNA was analyzed as a loading control. “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s012

(PDF)

S12 Fig. Overexpression of RNAi factors does not affect splicing of intron-containing centromeric ncRNAs.

A multicopy plasmid (pSP1) harboring the dcr1+ or cid12+ gene was introduced into wild-type cells, and splicing of dg, antisense dg and dh ncRNAs was analyzed by RT-PCR (panels of dg intron, antisense dg intron and dh intron). Overexpression of the dcr1+ and cid12+ genes was confirmed by RT-PCR (panels of dcr1 mRNA and cid12 mRNA). Act1 mRNA was analyzed as a loading control (act1 mRNA). “RT-” indicates a reaction without reverse transcriptase.

https://doi.org/10.1371/journal.pgen.1006606.s013

(PDF)

S13 Fig. Predicted structure of the dg intron region.

Secondary structure of the dg intron with upstream (-100 bp from the dg intron) and downstream (+100 bp from the dg intron) sequences was predicted using Mfold (http://unafold.rna.albany.edu/?q=mfold/RNA-Folding-Form). Mutations introduced in the splice sites or branch point shown in Fig 7 did not affect the predicted structure significantly. Because this region is included in the dg-short construct, deletions of the upstream sequence shown in Fig 8A also do not affect the structural prediction. The intron region is indicated in red.

https://doi.org/10.1371/journal.pgen.1006606.s014

(PDF)

S14 Fig. Predicted structure of the dh intron region.

Secondary structure of the dh intron with upstream (-50 bp from the dh intron) and downstream (+200 bp from the dh intron) sequences was predicted using Mfold (http://unafold.rna.albany.edu/?q=mfold/RNA-Folding-Form). The intron region is indicated in red.

https://doi.org/10.1371/journal.pgen.1006606.s015

(PDF)

Acknowledgments

We thank Dr. K. Gull for providing the TAT1 antibody, Dr. R.C. Allshire for strains with the ura4 insertion, Dr. D. Moazed for strains with Cid12-TAP Δrdp1, Dr. M. Baum for centromere plasmids, and the National Bio-Resource Project (NBRP), Japan, for providing the strain expressing Nuf2-RFP.

Author Contributions

  1. Conceptualization: TT MMu.
  2. Data curation: MMu TT.
  3. Formal analysis: MMu.
  4. Funding acquisition: TT MMu.
  5. Investigation: MMu MMo CT MC SN TY MS KD KI TI TT.
  6. Methodology: MMu MMo JiN MC KI TT.
  7. Project administration: TT.
  8. Resources: TT JiN KI.
  9. Supervision: TT.
  10. Validation: MMu MMo CT MC KD KI.
  11. Visualization: MMu TT.
  12. Writing – original draft: MMu TT.
  13. Writing – review & editing: MMu TT MC JiN KI.

References

  1. 1. Pidoux AL, Allshire RC (2004) Kinetochore and heterochromatin domains of the fission yeast centromere. Chromosome Res 12(6):521–534. Epub 2004/08/04. pmid:15289660
  2. 2. Verdel A, Jia S, Gerber S, Sugiyama T, Gygi S, Grewal SI, et al. (2004) RNAi-mediated targeting of heterochromatin by the RITS complex. Science 303(5658):672–676. Epub 2004/01/06. pmid:14704433
  3. 3. Verdel A, Moazed D (2005) RNAi-directed assembly of heterochromatin in fission yeast. FEBS Lett 579(26):5872–5878. pmid:16223485
  4. 4. White SA, Allshire RC (2008) RNAi-mediated chromatin silencing in fission yeast. Curr Top Microbiol Immunol 320:157–183. Epub 2008/02/14. pmid:18268844
  5. 5. Bayne EH, Portoso M, Kagansky A, Kos-Braun IC, Urano T, Ekwall K, et al. (2008) Splicing factors facilitate RNAi-directed silencing in fission yeast. Science 322(5901):602–606. Epub 2008/10/25. PubMed Central PMCID: PMC2585287. pmid:18948543
  6. 6. Chinen M, Morita M, Fukumura K, Tani T (2010) Involvement of the spliceosomal U4 small nuclear RNA in heterochromatic gene silencing at fission yeast centromeres. J Biol Chem 285(8):5630–5638. Epub 2009/12/19. PubMed Central PMCID: PMC2820790. pmid:20018856
  7. 7. Kallgren SP, Andrews S, Tadeo X, Hou H, Moresco JJ, Tu PG, et al. (2014) The proper splicing of RNAi factors is critical for pericentric heterochromatin assembly in fission yeast. PLoS Genet 10(5):e1004334. PubMed Central PMCID: PMCPMC4038458. pmid:24874881
  8. 8. Urushiyama S, Tani T, Ohshima Y (1996) Isolation of novel pre-mRNA splicing mutants of Schizosaccharomyces pombe. Mol Gen Genet 253(1–2):118–127. Epub 1996/11/27. pmid:9003295
  9. 9. Cordin O, Beggs JD (2013) RNA helicases in splicing. RNA Biol 10(1):83–95. Epub 2012/12/12. PubMed Central PMCID: PMC3590240. pmid:23229095
  10. 10. Motamedi MR, Verdel A, Colmenares SU, Gerber SA, Gygi SP, Moazed D (2004) Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. Cell 119(6):789–802. Epub 2004/12/21. pmid:15607976
  11. 11. Baum M, Ngan VK, Clarke L (1994) The centromeric K-type repeat and the central core are together sufficient to establish a functional Schizosaccharomyces pombe centromere. Mol Biol Cell 5(7):747–761. PubMed Central PMCID: PMCPMC301093. pmid:7812044
  12. 12. Grewal SI, Jia S (2007) Heterochromatin revisited. Nat Rev Genet 8(1):35–46. Epub 2006/12/19. pmid:17173056
  13. 13. Almouzni G, Probst AV (2011) Heterochromatin maintenance and establishment: lessons from the mouse pericentromere. Nucleus 2(5):332–338. pmid:21941119
  14. 14. Wood V, Gwilliam R, Rajandream MA, Lyne M, Lyne R, Stewart A, et al. (2002) The genome sequence of Schizosaccharomyces pombe. Nature 415(6874):871–880. pmid:11859360
  15. 15. Prabhala G, Rosenberg GH, Kaufer NF (1992) Architectural features of pre-mRNA introns in the fission yeast Schizosaccharomyces pombe. Yeast 8(3):171–182. pmid:1574925
  16. 16. Lee NN, Chalamcharla VR, Reyes-Turcu F, Mehta S, Zofall M, Balachandran V, et al. (2013) Mtr4-like protein coordinates nuclear RNA processing for heterochromatin assembly and for telomere maintenance. Cell 155(5):1061–1074. Epub 2013/11/12. PubMed Central PMCID: PMC3974623. pmid:24210919
  17. 17. Schwer B, Guthrie C (1991) PRP16 is an RNA-dependent ATPase that interacts transiently with the spliceosome. Nature 349(6309):494–499. Epub 1991/02/07. pmid:1825134
  18. 18. Burge CB, Tuschl T, Sharp PA. Splicing of precursors to mRNA by the spliceosome. In: Gesteland RF, Cech TR, Atkins JF, editors. The RNA word. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 1999. p. 525–560.
  19. 19. Bernard P, Drogat J, Dheur S, Genier S, Javerzat JP (2010) Splicing factor Spf30 assists exosome-mediated gene silencing in fission yeast. Mol Cell Biol 30(5):1145–1157. Epub 2009/12/24. PubMed Central PMCID: PMC2820897. pmid:20028739
  20. 20. Dumesic PA, Natarajan P, Chen C, Drinnenberg IA, Schiller BJ, Thompson J, et al. (2013) Stalled spliceosomes are a signal for RNAi-mediated genome defense. Cell 152(5):957–968. PubMed Central PMCID: PMCPMC3645481. pmid:23415457
  21. 21. Colot HV, Stutz F, Rosbash M (1996) The yeast splicing factor Mud13p is a commitment complex component and corresponds to CBP20, the small subunit of the nuclear cap-binding complex. Genes Dev 10(13):1699–1708. pmid:8682299
  22. 22. Lewis JD, Izaurralde E, Jarmolowski A, McGuigan C, Mattaj IW (1996) A nuclear cap-binding complex facilitates association of U1 snRNP with the cap-proximal 5' splice site. Genes Dev 10(13):1683–1698. pmid:8682298
  23. 23. Rymond BC, Rosbash M. Yeast pre-mRNA splicing. In: Jones EW, Pringle JR, Broach JR, editors. The molecular and cellular biology of the yeast Saccharomyces cerevisiae. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1992. p. 143–192.
  24. 24. Djupedal I, Kos-Braun IC, Mosher RA, Soderholm N, Simmer F, Hardcastle TJ, et al. (2009) Analysis of small RNA in fission yeast; centromeric siRNAs are potentially generated through a structured RNA. EMBO J 28(24):3832–3844. PubMed Central PMCID: PMCPMC2797062. pmid:19942857
  25. 25. Zuker M (2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31(13):3406–3415. PubMed Central PMCID: PMCPMC169194. pmid:12824337
  26. 26. Izuta H, Ikeno M, Suzuki N, Tomonaga T, Nozaki N, Obuse C, et al. (2006) Comprehensive analysis of the ICEN (Interphase Centromere Complex) components enriched in the CENP-A chromatin of human cells. Genes Cells 11(6):673–684. Epub 2006/05/24. pmid:16716197
  27. 27. Ideue T, Cho Y, Nishimura K, Tani T (2014) Involvement of satellite I noncoding RNA in regulation of chromosome segregation. Genes Cells 19(6):528–538. pmid:24750444
  28. 28. Moreno S, Klar A, Nurse P (1991) Molecular genetic analysis of fission yeast Schisosaccharomyces pombe. Methods Enzymology 194:795–823.
  29. 29. Alfa C, Fantes P, Hyams J, McLeod M, Warbrick E. Experiments with fission yeast. New York: Cold Spring Harbor Laboratory Press; 1993.
  30. 30. Folco HD, Pidoux AL, Urano T, Allshire RC (2008) Heterochromatin and RNAi are required to establish CENP-A chromatin at centromeres. Science 319(5859):94–97. PubMed Central PMCID: PMCPMC2586718. pmid:18174443