Meiotic gene silencing complex MTREC/NURS recruits the nuclear exosome to YTH-RNA-binding protein Mmi1

Summary

Accurate target recognition in transcript degradation is crucial for regulation of gene expression. In the fission yeast Schizosaccharomyces pombe, a number of meiotic transcripts are recognized by a YTH-family RNA-binding protein, Mmi1, and selectively degraded by the nuclear exosome during mitotic growth. Mmi1 forms nuclear foci in mitotically growing cells, and the nuclear exosome colocalizes to such foci. However, it remains elusive how Mmi1 and the nuclear exosome are connected. Here, we show that a complex called MTREC (Mtl1-Red1 core) or NURS (nuclear RNA silencing) that consists of a zinc-finger protein, Red1, and an RNA helicase, Mtl1, is required for the recruitment of the nuclear exosome to Mmi1 foci. Physical interaction between Mmi1 and the nuclear exosome depends on Red1. Furthermore, a chimeric protein involving Mmi1 and Rrp6, which is a nuclear-specific component of the exosome, suppresses the ectopic expression phenotype of meiotic transcripts in red1Δ cells and mtl1 mutant cells. These data indicate that the primary function of MTREC/NURS in meiotic transcript elimination is to link Mmi1 to the nuclear exosome physically.

Keywords:

Gene expression – Messenger RNA – Schizosaccharomyces pombe – Plasmid construction – Immunoprecipitation – Yellow fluorescent protein – Co-immunoprecipitation – Exosomes

Introduction

Stability control of transcripts is crucial to establish appropriate gene expression profiles in a wide range of biological processes [1, 2]. In the fission yeast Schizosaccharomyces pombe, expression of a number of meiotic genes is strictly repressed during the mitotic cell cycle by selective RNA elimination, in addition to transcriptional suppression, since their aberrant expression is highly deleterious to cell growth [3, 4]. Selectivity in meiotic RNA degradation is guaranteed by a particular sequence element termed DSR (determinant of selective removal) on meiotic transcripts, which is composed of the repetitive hexanucleotide motif UNAAAC [3, 5–7]. Meiotic transcripts are recognized by a YTH family RNA-binding protein, Mmi1, through direct binding of the YTH domain with UNAAAC DSR motifs [3, 6]. The mode of Mmi1 interaction with RNAs is different from that of other YTH proteins, and Mmi1 does not recognize N⁶-methyladenosine-containing RNAs [8–11].

Mmi1 induces selective degradation of DSR-containing meiotic transcripts by the nuclear exosome [3, 12]. The exosome is a highly conserved 3′-5′ exonuclease complex that localizes to both the nucleus and cytoplasm [13–15]. The cytoplasmic exosome consists of a nine-subunit inert complex and a catalytically active subunit, Dis3. In the nucleus, an additional exonuclease, Rrp6, binds to the nine-subunit complex in addition to Dis3. Recent genome-wide analyses have revealed that the exosome targets various classes of coding and non-coding transcripts [16–18]. The exosome cooperates with a variety of adapter complexes for selective degradation of divergent target transcripts. Mmi1-mediated selective elimination of meiotic transcripts requires polyadenylation-related factors, including a canonical poly(A) polymerase, Pla1, and a poly(A)-binding protein, Pab2 [12, 19, 20]. Another pivotal factor engaged in Mmi1-mediated RNA degradation is a zinc-finger protein, Red1 [21, 22]. Red1, together with Mtr4-like RNA helicase protein, Mtl1, constitutes a complex termed MTREC (Mtl1-Red1 core) or NURS (nuclear RNA silencing), and mediates the selective degradation of various transcripts, including not only meiotic RNAs, but also cryptic unstable transcripts (CUTs), the latter of which are independent of Mmi1 [23–25]. The machinery of Mmi1-mediated RNA degradation also triggers facultative heterochromatin formation at a subset of its target genes [5, 26, 27]. Furthermore, it regulates transcription termination of its targets [28–30], suggesting that Mmi1-mediated multilayered regulation rigorously shapes appropriate gene expression profiles in mitotically growing cells.

Mmi1 forms nuclear foci in vegetatively growing cells [3]. Most, if not all, factors cooperating with Mmi1, including the nuclear exosome, colocalize with Mmi1 at nuclear foci [12, 21, 22, 24, 31, 32]. We have recently shown that DSR-containing meiotic transcripts are tethered to the Mmi1 foci [33]. Tethering of meiotic transcripts leads to prevention of their nuclear export and deleterious expression of meiotic proteins, even when RNA elimination is compromised [33]. Foci formation of Mmi1 requires its self-interaction with the assistance of a homolog of Enhancer of Rudimentary, Erh1, which induces Mmi1 dimerization [33, 34]. Mmi1 mutant cells lacking the self-interaction domain (SID) and erh1 deletion mutant cells do not form Mmi1 foci. In these mutant cells, DSR-containing transcripts are liberated from the nuclear foci and escape degradation. These findings highlight the importance of spatial control in Mmi1-mediated regulation. However, it remains largely unknown how factors involved in Mmi1-mediated regulation are incorporated into the nuclear foci. The Red1-containing complex MTREC/NURS binds to both Mmi1 and the nuclear exosome [23–25], suggesting that Red1 acts as a bridge between the nuclear exosome and Mmi1. However, Red1 function in Mmi1-mediated transcript elimination remains to be elucidated.

Here, we show that recruitment of the nuclear exosome to nuclear foci is Red1-dependent. We further demonstrated that Red1 mediates physical interactions between Mmi1 and the nuclear exosome. Our current study sheds new light on our understanding of accurate target recognition in nuclear RNA degradation.

Results

Red1 is crucial for nuclear exosome foci formation

To elucidate the mechanisms underlying exosome foci formation, we searched for factors required for the proper localization of Rrp6. We first examined the effect of deletion of the red1, mmi1, and pab2 genes, whose gene products colocalize with nuclear foci of Rrp6 [12, 21]. Deletion of mmi1 causes severe growth defects due to the ectopic expression of meiotic transcripts. Growth defects of mmi1 deletion can be alleviated by deletion of mei4, a crucial target of Mmi1 [3]. Thus, we used a mmi1Δ mei4Δ double mutant for examining the impact of mmi1 deletion. Wild-type cells showed nucleolar localization of Rrp6. Rrp6 also formed foci in both nucleoplasm and nucleolus (Fig 1A and S1A Fig). Strikingly, nuclear foci formation of Rrp6 was severely impaired in red1Δ cells, whereas nucleolar accumulation was maintained. Cells lacking mmi1 exhibited only a modest reduction in the frequency of Rrp6 foci formation and the number of foci per cell, indicating weak contribution of Mmi1 in Rrp6 foci formation. Foci formation of Mmi1 and Red1 was not interdependent (Fig 1A and S1A Fig) [22, 33], and they might play divergent roles in Rrp6 foci formation. The localization of Rrp6 in pab2Δ cells was comparable to that observed in wild-type cells. We also found that deletion of red1 strongly impeded foci formation of both Dis3 and Rrp4, the core components of the exosome (Fig 1B and S1B Fig). Since the fluorescence signal of Dis3 and Rrp4 in the nucleolus in wild-type cells was more intense than that of Rrp6, the frequency of foci formation in or on the border of the nucleolus, and in nucleoplasm, might be underestimated. Deletion of red1 had no severe impact on nuclear foci formation of Pla1 and Pab2, although percentages of cells containing 1 focus were increased (Fig 1C and 1D, S1C Fig and S1D Fig). From these observations, we concluded that Red1 is a key player in nuclear foci formation of the exosome and regulates the localization of a subset of factors involved in Mmi1-mediated RNA degradation.

Rrp6 foci formation is dependent on an N-terminal domain of Red1

We next determined which region of Red1 was essential for nuclear foci formation. Since it has been reported that the C-terminal zinc-finger motif is required for the elimination of meiotic transcripts but not for localization [21], we constructed N-terminal deletion mutants according to secondary structure predictions by JPred (Fig 2A) [35]. Comparable expression of the Red1 deletion series was confirmed by western blot analysis (S2A Fig). The deletion of residues 2 to 195 did not alter the localization of Red1 (Fig 2A and 2B). When residues 2 to 347 was deleted, Red1 failed to form nuclear foci and was, rather, uniformly localized in the nucleus. The deletion of residues 196 to 347 also resulted in uniform nuclear localization. We then divided the domain involving residues 196 to 347 into three regions, and found that residues 196 to 245 were responsible for nuclear foci formation of Red1.

We next examined Rrp6 localization in cells expressing Red1 lacking residues 196 to 245 (red1(Δ196–245)), and found that this domain is vital for Rrp6 nuclear foci formation (Fig 2C and S2B Fig). We also tested the effect of deletion of iss10, which is essential for Red1 nuclear foci formation [22, 24]. iss10Δ cells exhibited a strong reduction in Rrp6 foci formation frequency (Fig 2D and S2C Fig), suggesting that the proper localization of Red1 is crucial for nuclear foci formation of Rrp6.

The growth profile of red1(Δ196–245) cells was similar to that of wild-type cells, while red1Δ cells grew more slowly than wild-type cells and exhibited severe cold sensitivity (S3A Fig) [21, 22]. iss10Δ cells showed mild cold sensitivity, as previously demonstrated [22]. The cold sensitivity of red1Δ is not suppressed by deletion of mei4, indicating that the defect arises independently of ectopic accumulation of meiotic transcripts [32, 33]. Meiotic transcripts, such as mei4, ssm4 and rec8 mRNAs, were accumulated in red1(Δ196–245) and iss10Δ cells during vegetative growth, albeit to a lesser extent than red1Δ cells (Fig 2E and S3B Fig). The effect of the deletion of residues 196 to 245 was less prominent on expression of spo5 mRNAs, which are also targeted by Mmi1. These results suggest that Red1 has some activity without this region or in the absence of Iss10 and that residues 196 to 245 is required to fully induce meiotic transcript degradation.

Red1 has been shown to mediate degradation of CUTs in an Mmi1-independent manner [25]. We then examined the expression of the promoter upstream transcripts (PROMPTs) of the cti6 gene and the rpl402 gene, which are known to be CUTs transcribed from the promoter region of cti6 or rpl402 in anti-sense direction. In contrast to meiotic transcripts such as ssm4, rec8 and spo5, ectopic expression of the cti6 and rpl402 PROMPT was detected in red1Δ and rrp6Δ cells, but not in mmi1Δ cells (S3C Fig), as shown previously [25]. red1(Δ196–245) and iss10Δ cells showed a modest, if any, defect in elimination of the cti6 and rpl402 PROMPT, as was the case with meiotic transcripts (S3D Fig). Altogether, the 196–245 domain of Red1 was required for nuclear foci formation of itself and Rrp6, and played a significant role in regulation of meiotic transcript and CUT expression, although it was dispensable for growth at low temperatures.

Red1 connects Rrp6 with Mmi1 for recruiting target RNAs to the exosome

Red1 physically interacts with both Rrp6 and Mmi1 [21]. To test whether Red1 mediates the interaction between Mmi1 and Rrp6, we carried out co-immunoprecipitation assays. Mmi1 was specifically co-purified with Rrp6 (S4A Fig). RNase treatment did not affect the interaction of Mmi1 with Rrp6 (S4B Fig), indicating that this interaction is independent of RNA molecules. The interaction between these proteins was abolished in the absence of red1 (Fig 3A), suggesting that Red1 physically links Mmi1 with the exosome. The deletion of residues 196 to 245 in Red1 mildly impaired the interaction between Mmi1 and Rrp6 owing to reduction of the interaction of Red1 with Mmi1 (Fig 3B).

The above results implied that gene expression defects in red1Δ cells could be ascribed to impaired interaction between Mmi1 and Rrp6. We investigated this possibility by expressing a chimeric Mmi1-Rrp6 protein in red1Δ cells. We constructed a plasmid carrying a chimeric gene in which GFP was inserted between full-length rrp6 and mmi1 (rrp6-GFP-mmi1) ORFs. The chimeric protein Rrp6-GFP-Mmi1 was able to suppress the growth defect, and ectopic accumulation of meiotic transcripts and CUTs in rrp6Δ cells (S5A and S5B Fig). Rrp6-GFP-Mmi1 also suppressed the growth defect and ectopic accumulation of meiotic transcripts in temperature-sensitive mmi1 mutant cells (mmi1-ts3) (S5C and S5D Fig). In these experiments, cells were cultured in minimal medium to induce expression of the chimeric protein. Since mmi1Δ cells show growth retardation in minimal medium even when combined with the mei4 deletion, we used the temperature-sensitive mmi1 mutant instead of mmi1Δ. These data indicated that Rrp6-GFP-Mmi1 maintains the functions of both Rrp6 and Mmi1. Intriguingly, aberrant accumulation of mei4, ssm4, rec8 and spo5 mRNAs in red1Δ cells was abrogated when Rrp6-GFP-Mmi1 was expressed, whereas not by either Rrp6 or Mmi1 alone (Fig 4A and S6A Fig). It should be noted that the use of minimal medium in these experiments resulted in the lower accumulation of meiotic transcripts than that in rich medium (S3 Fig). Meanwhile, cti6 and rpl402 PROMPTs still accumulated in red1Δ cells even when Rrp6-GFP-Mmi1 was expressed (Fig 4A and S6A Fig). Rrp6-GFP-Mmi1 did not suppress the cold sensitivity of red1Δ cells (S6B Fig), consistently with that the growth defect caused by red1 deletion is irrelevant to ectopic expression of meiotic transcripts [32, 33]. We found that the slight growth retardation of red1Δ cells at 25 to 36˚C, which has been shown previously [21, 22], was suppressed by expression of Rrp6, Mmi1 or Rrp6-GFP-Mmi1 as well as Red1, although the underlying mechanism remains unclear.

We next confirmed the Red1-mediated interaction between Mmi1 and Rrp6 by in vitro binding assay (Fig 4B). The weak direct interaction of Mmi1 with Rrp6 was observed in our experimental conditions. Addition of Red1 greatly increased the interaction, indicating the role of Red1 in connecting Mmi1 to the exosome. The direct binding of Red1 with Rrp6 was also observed (Fig 4B).

From these results, we concluded that the physical interaction between Rrp6 and Mmi1 is crucial for the elimination of meiotic transcripts, but not for other function(s) involving Red1.

Self-interaction of Mmi1 is vital for meiotic transcript elimination by Rrp6

We previously demonstrated that Mmi1 interacts with itself via the self-interaction domain, SID, which is essential for the proper function and localization of Mmi1 [33]. Given that Rrp6-GFP-Mmi1 functions by interacting with endogenously expressed Mmi1, the YTH-RNA-binding domain in the chimeric protein could be unnecessary for meiotic transcript elimination because target transcripts might be recognized by endogenous Mmi1 and delivered to Rrp6 in the chimeric protein through Mmi1-Mmi1 interaction. To test this hypothesis, we constructed a plasmid carrying a fusion gene of rrp6, GFP or YFP, and mmi1-ΔYTH or mmi1-ΔSID, lacking the region encoding the YTH-RNA-binding domain or the SID, respectively. We confirmed that chimeric proteins were expressed in a comparable amount, although removal of the SID resulted in an increase of the expression level (S6C Fig). Both chimeric proteins maintained the function of Rrp6 (S5A and S5B Fig), but lost the ability to act as Mmi1 (S5C and S5D Fig). The chimeric protein containing Mmi1-ΔYTH showed a negative effect in mmi1-ts3 cells at permissive temperatures (S5C and S5D Fig). This is because Mmi1 lacking the YTH domain exerts a dominant-negative activity, as shown previously [33]. Rrp6-GFP-Mmi1-ΔYTH amended the degradation defect of mei4, ssm4, rec8 and spo5 mRNAs in red1Δ cells, but Rrp6-GFP-Mmi1-ΔSID did not (Fig 4A and S6A Fig). Rrp6-GFP-Mmi1-ΔYTH did not suppress the degradation defect in red1Δ mmi1Δ cells (Fig 4C and S6D Fig), indicating that Rrp6-GFP-Mmi1-ΔYTH requires endogenous Mmi1 to exert the function in the absence of Red1. The interaction between Rrp6-GFP-Mmi1 and endogenous Mmi1 was demonstrated by co-immunoprecipitation (S6E Fig). This interaction was dependent on the SID but not on the YTH domain, as predicted.

We next examined the localization of the chimeric proteins. YFP instead of GFP was used for the observation to accommodate the microscope filter set. Rrp6-YFP-Mmi1 formed nuclear foci in red1Δ cells (Fig 4D). The absence of the YTH domain did not affect the localization of the chimeric protein. However, SID deletion significantly reduced the frequency of foci formation. These observations suggested that Mmi1 self-interaction is crucial for foci formation of the chimeric proteins, and for meiotic transcript degradation by Rrp6, although we cannot exclude the possibility that the SID is required for the interaction of Mmi1 with other factors required for meiotic RNA elimination.

Rrp6 is required for foci formation of other exosome components

Rrp6 mediates the interaction between core components of the exosome and various exosome-related factors [36]. To examine whether Rrp6 acts as a mediator between Red1 and other exosome components, we observed the localization of core components of the exosome in rrp6Δ cells. Dis3 and Rrp4 failed to form foci in the absence of rrp6 (Fig 5A and S7A Fig), while Red1 and Mmi1 showed similar localization to wild-type cells (Fig 5B and S7B Fig). In red1Δ cells, in which foci formation of Rrp6, Dis3 and Rrp4 was severely impaired (Fig 1A and 1B), Rrp6-YFP-Mmi1 partially recovered the foci formation frequencies of Dis3 and Rrp4 (Fig 5C and S7C Fig). Foci formation frequency was similar even when the YTH domain was absent. The recovery of Dis3 and Rrp4 foci formation depended on the SID of Mmi1 in the chimeric protein. These results suggested that Rrp6 plays a key role in the assembly of the exosome at nuclear foci containing Mmi1 and Red1.

Red1 links Mmi1 to the exosome in cooperation with Mtl1

We next investigated whether the function of Red1 to connect Mmi1 with the exosome required Mtl1, a component of the Red1-containing MTREC/NURS complex [23, 24]. Nuclear foci formation of Rrp6 was compromised in mtl1 mutant cells (Fig 6A). The relatively mild effect compared to that in red1Δ cells (Fig 1A) might arise from the use of a conditional cold-sensitive mutant, since Mtl1 is essential for cell growth. We further examined the expression of meiotic transcripts in the mtl1-cs5 cells expressing Rrp6-GFP-Mmi1. Expression of the chimeric protein suppressed aberrant accumulation of meiotic mRNAs as efficiently as that of wild-type Mtl1 (Fig 6B and S8 Fig). These observations suggest that Mtl1 contributes to connect Mmi1 and the nuclear exosome by forming the MTREC/NURS complex with Red1.

Discussion

In this study, we demonstrated that the primary function of Red1 in meiotic transcript elimination is to connect Mmi1 and the nuclear exosome physically. We have also shown that Rrp6 is crucial for nuclear foci formation of the core exosome. This observation is consistent with the result of previous co-purification assays, which showed that the interaction between exosome subunits and Red1 depends on Rrp6 [25]. Red1 forms the complex, MTREC or NURS, which includes an RNA helicase, Mtl1 [23, 24]. Our observations also imply that the MTREC/NURS complex, rather than Red1 alone, may link Mmi1 to the exosome.

We found that foci formation of Rrp6 was severely impaired by the deletion of residues 196 to 245 in Red1. Meanwhile, ectopic expression levels of meiotic transcripts were moderate in red1(Δ196–245) cells compared to those in red1Δ cells. These results suggest that nuclear foci, to which Rrp6 and Mmi1 localize, might not be the exclusive sites for Mmi1-mediated RNA degradation. Indeed, the localization of factors involved in Mmi1-mediated degradation is not limited to nuclear foci; Mmi1, Red1 and Rrp6 have been shown to localize to the loci encoding target transcripts including mei4 by chromatin immunoprecipitation experiments [23, 24, 26, 27, 32, 37], although their colocalization with the mei4 locus was not observed by microscopic observations [24, 37]. Thus, it is plausible that Mmi1-mediated RNA degradation also takes place at the genetic loci where its targets are transcribed. Further clarification of the site(s) of RNA degradation would be important for understanding of Mmi1-mediated regulation.

Red1 exerts various functions beyond meiotic transcript elimination. For instance, Red1 plays an essential role in degradation of CUTs by the nuclear exosome, independently of Mmi1 [25]. Red1 is also required for growth at low temperatures [21, 22], although it remains elusive what Red1 does under cold conditions. It is interesting whether Red1 physically links a target-recognition factor to the degradation machinery, such as the nuclear exosome, in other situations as in meiotic transcript elimination. Another intriguing question is whether Red1 function is conserved in other organisms. A human zinc-finger protein, ZFC3H1, which is suggested to be a counterpart of Red1, forms a complex termed PAXT (poly(A) tail exosome targeting) together with an Mtl1-ortholog, hMtr4, and plays an important role in the selective elimination of polyadenylated nuclear RNAs [38, 39]. Intriguingly, ZFC3H1 is required not only for RNA degradation but also retention of target transcripts to nuclear foci [39, 40]. We have demonstrated that Red1 is dispensable for tethering of meiotic transcripts to nuclear foci, while Mmi1 prevents nuclear export of its targets by sequestering them to foci even when RNA degradation is dampened [33]. Future studies may clarify the conservation of Red1 function in higher eukaryotes.

Materials and Methods

Yeast strains and general genetic methods

The S. pombe strains used in this study are listed in S1 Table. The general genetic manipulation procedures and growth media have been previously described [41, 42]. Deletion mutants and epitope-tagged strains were constructed using PCR-based gene-targeting protocols [43, 44].

To construct truncated red1 mutant strains, we cloned the red1 open reading frame (ORF) with 1 kb upstream and downstream flanking regions and removed the indicated regions using the PrimeStar Mutagenesis kit (Takara, Shiga, Japan). PCR-amplified DNA fragments encompassing each truncated red1 ORF with flanking regions were introduced into the red1::ura4⁺ strain. Transformants were counter-selected on medium containing 5-fluoroorotic acid.

Plasmids expressing chimeric rrp6-GFP-mmi1 gene and its variants were constructed by cloning the PCR-amplified rrp6 and full-length or truncated mmi1 ORFs into pREP81 carrying the green fluorescent protein (GFP) ORF, in which the fusion genes were expressed under the control of a modified thiamine-repressive nmt1 promoter [45]. To observe the localization and expression levels of chimeric proteins (Figs 4D and 5C, S6C Fig and S7C Fig), GFP was replaced with yellow fluorescent protein (YFP) and the fusion genes were expressed from the modified constitutively active adh1 promoter.

Fluorescence microscopy

Cells were grown in logarithmic phase at 30°C and mounted onto lectin-coated glass bottom culture dishes (MatTek, Ashland, MA) filled with liquid growth medium. Images were acquired using the DeltaVision-SoftWoRx system (GE Healthcare, Chicago, IL) by collecting 12 optical sections along the z-axis at 0.5 μm intervals. All images were deconvolved and merged into single projections using SoftWoRx software.

RNA preparation and reverse-transcription quantitative PCR (RT-qPCR) analysis

Total RNA extraction and RT-qPCR analysis were performed as previously described [33] by using ReverTra Ace qPCR Master Mix (TOYOBO, Osaka, Japan) and a LightCycler96 instrument (Roche, Basel, Switzerland) with SYBR Premix Ex Taq II (Takara). Normalization was performed using act1, which encodes actin. Primers used in this study are listed in S2 Table.

Immunoprecipitation and western blot analysis

Immunoprecipitation was performed as previously described [46]. Hemagglutinin (HA)-tagged Rrp6 was precipitated using mouse anti-HA-tag magnetic beads M180-11 (MBL, Nagoya, Japan). Red1-YFP and GFP-containing chimeric proteins were precipitated and detected using anti-GFP-tag magnetic beads D153-11 (MBL) and anti-GFP antibody [47], respectively. Rabbit anti-Mmi1 TB0514 (our laboratory preparation) and mouse anti-HA 12CA5 (Sigma-Aldrich, St. Louis, MO) were used to detect Mmi1 and Rrp6-3HA, respectively. For immunoprecipitation with RNase treatment, cell lysates were incubated for 30 min at room temperature with 0.5 U of RNase A and 20 U of RNase T1 (RNase Cocktail Enzyme Mix, Thermo Fisher Scientific, Waltham, MA). In S2A Fig and S6C Fig, harvested cells were disrupted with glass beads in 20% trichloroacetic acid. Anti-GFP antibody [47] was used to detect Red1-YFP, Rrp6-YFP and Rrp6-YFP-Mmi1. Anti-γ-tubulin GTU-88 (Sigma-Aldrich) was used for a loading control.

in vitro binding assay

PCR-amplified YFP-mmi1 and rrp6-FLAG were cloned into pCold-GST (Takara). Glutathione S-transferase (GST)-tagged YFP-Mmi1 and Rrp6-FLAG proteins purified from E. coli were separated by NGC Chromatography System (Bio-rad, Hercules, CA) with HiTrap Heparin HP and HiTrap Q HP column (GE Healthcare) for GST-YFP-Mmi1 and GST-Rrp6-FLAG, respectively. It should be noted that most of GST-Rrp6-FLAG was cleaved between GST and Rrp6-FLAG during the preparation. HA-tagged Red1 was expressed in Rabbit Reticulocyte Lysate System, Nuclease Treated (Promega, Madison, WI) according to the manufacturer’s instruction. mRNA containing the His-tagged red1-3HA gene was transcribed using T7-Scribe Standard RNA IVT Kit (Cellscript, Madison, WI). Capping and poly(A)-tailing were performed using ScriptCap m⁷G Capping System (Cellscript), ScriptCap 2′-O-Methyltransferase Kit (Cellscript), and A-Plus Poly(A) Polymerase Tailing Kit (Cellscript). For negative controls, we parallelly performed in vitro translation without mRNA. Proteins were incubated at the room temperature for 30 min in the binding buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 5 mM MgCl₂, 1 mM dithiothreitol (DTT), 0.5% Triton-X100, 5% glycerol), and then mixed with Anti-DDDDK-tag mAb-Magnetic Beads M185-11 (MBL) at 4°C for 30 min. Beads were washed by wash buffer (20 mM HEPES pH 7.5, 300 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 1% Triton-X100) for 5 times, and subjected to western blotting. Rabbit anti-GFP ab6556 (abcam, Cambridge, UK), mouse anti-FLAG M2 (Sigma-Aldrich), and rat anti-HA 3F10 (Sigma-Aldrich) were used to detect GST-YFP-Mmi1, Rrp6-FLAG, and Red1-3HA, respectively. Images were acquired by Odyssey CLx Infrared Imaging System (LI-COR, Lincoln, NE).

Supporting information

S1 Fig [a]
Red1 is required for nuclear foci formation of exosome.

S2 Fig [a]
Red1(Δ196–245) is defective in Rrp6 foci formation.

S3 Fig [a]
Red1(Δ196–245) is defective in meiotic transcript degradation.

S4 Fig [a]
Rrp6 physically interacts with Mmi1.

S5 Fig [a]
Chimeric Rrp6-GFP-Mmi1 protein maintains the function of Rrp6 and Mmi1.

S6 Fig [a]
Interaction between Rrp6 and Mmi1 is essential for meiotic transcript elimination.

S7 Fig [a]
Rrp6 is vital for nuclear foci formation of other exosome components.

S8 Fig [pdf]
Mtl1 mediates interaction between Rrp6 and Mmi1.

S1 Table [pdf]
Strains used in this study.

S2 Table [pdf]
Primers used in this study.

S3 Table [xlsx]
Numerical data for all of graphs in this study.