ABSTRACT
Stress granules (SGs) are cytoplasmic aggregates formed upon stress when untranslated messenger ribonucleoproteins accumulate in the cells. In a green fluorescent protein library screening of the fission yeast SG proteins, Puf2 of the PUF family of RNA-binding proteins was identified that is required for SG formation after deprivation of glucose. Accordingly, the puf2 mutant is defective in recovery from glucose starvation with a much longer lag to reenter the cell cycle. In keeping with these results, Puf2 contains several low-complexity and intrinsically disordered protein regions with a tendency to form aggregates and, when overexpressed, it represses translation to induce aggregation of poly(A) binding protein Pabp, the signature constituent of SGs. Intriguingly, overexpression of Puf2 also enhances the structure of processing bodies (PBs), another type of cytoplasmic RNA granule, a complex of factors involved in mRNA degradation. In this study, we demonstrate a function of the fission yeast PB in SG formation and show Puf2 may provide a link between these two structures.
INTRODUCTION
RNA granules are microscopically visible non-membrane-bound cellular structures that consist of condensates of mRNAs bound to a plethora of proteins involved in translation and RNA degradation (1). Their structures are believed to form by dynamic reversible protein condensation or aggregation to control the localization, stability, and translation of their RNA cargo. Examples of RNA granules associated with mRNA metabolism include stress granules (SGs) and processing bodies (PBs). Their classification is based mostly on their localization, composition, and proposed function. Different types of RNA granules, such as SGs and PBs, contain many unique components with unique functions but also contain some common protein constituents using similar mechanisms to regulate mRNA translation and/or decay. To gain more insight into the biological function and regulation of SGs and PBs, we used the fission yeast Schizosaccharomyces pombe as a model system to study these structures.
PBs are cytoplasmic RNA granules that contain translationally repressed mRNAs and mRNA decay factors, including Dcp2, Dcp1, Edc3, Pat1, Lsm1-7, and Dhh1. Functions of PBs have been implicated in mRNA degradation, nonsense-mediated mRNA decay, translational repression, and RNA-mediated gene silencing (2). Because of the important role of PBs in diverse cellular functions, the assembly of these structures has been investigated in several model systems, and studies in the budding yeast Saccharomyces cerevisiae have been crucial in unravelling PB biology (3). However, Edc4, the third component of the decapping enzyme that appears to be the central component of decapping complex in high eukaryote, has no obvious homolog in budding yeast. The different protein composition of PBs between different species was contrasted by our study with the identification of the fission yeast Pdc1 as the functional homologue of Edc4, which is thought to be absent from fungi (4). Pdc1 forms a complex with Dcp2. Similar to Edc4, Pdc1 is an enhancer of decapping and plays a vital role in the formation of PBs. Our study of the characterization of the fission yeast Pdc2/Pat1 protein further reveals that even though species sharing similar protein composition, the underlying molecular mechanisms for their functions can be distinct (5). The whole complex of the decapping enzyme and its coactivators might have coevolved together and acquired additional proteins and mechanisms for its function.
In response to environmental stress, eukaryotic cells shut down protein synthesis in a stereotypic response that conserves anabolic energy for the repair of stress-induced damage. This results in the disassembly of polysomes leading to stalled initiation complexes, which are dynamically recruited to cytoplasmic foci called SGs (1). Accordingly, SG formation modulates the stress responses, viral infection, and signaling pathways (6, 7). Persistent or aberrant SG formation contributes to neurodegenerative disease in humans (8). SGs and PBs can dock and/or overlap each other in both budding yeast and mammalian cells, suggesting a dynamic mRNA cycle wherein messenger ribonucleoproteins can be remodeled and exchanged between these assemblies (9). SGs have been primarily studied in mammalian cells. The existence of SGs in the fission yeast S. pombe (10) and the distantly related budding yeast S. cerevisiae (11) was demonstrated more recently. These organisms contain many proteins analogous to those in mammalian SGs. Despite the fact that yeast SGs seem to contain most if not all components of mammalian SGs; unlike the situations in mammals, their formation is independent of eukaryotic initiation factor 2 α subunit (eIF2α) phosphorylation in yeast and in trypanosomes (10, 12, 13). It appears that different pathways contribute to the formation of SGs in different organisms (14), which warranted for further investigation. To this end, in this study we described a green fluorescent protein (GFP) fusion library screening for the fission yeast SG proteins. Several interesting aspects with regard to the structure of SGs were observed. In addition, members of the PUF family of RNA-binding proteins, including Puf1, -2, -3, and -4, were identified as components of the fission yeast SGs.
The PUF (Pumilio and FBF) family of RNA-binding proteins is known for its roles in cell division, differentiation, and development. All PUFs contain a PUM-HD-type RNA-binding domain, which folds into an arc-like shape with the capacity of binding to RNA and protein (15). The best-characterized function of PUFs is as posttranscriptional repressors (16). PUFs bind to specific recognition sequence in the 3′ untranslated region of mRNA to control the translation and the stability of the transcript. Recent studies have indicated that PUFs can also activate gene expression. Moreover, it is becoming clear that PUFs facilitate mRNA localization for spatial control of expression. Although a role for Pum2, a human PUF, in SGs has been suggested (17), a detailed study of the function of PUFs in SGs has not been described. Our study suggests that Puf2 repressed translation to induce SG formation and promoted protein aggregation through its low-complexity and intrinsically disordered protein regions. Intriguingly, we found that Puf2, when overexpressed, also affected the structures of PBs. A role for Puf2 in PB function was suggested, and a connection between SGs and PBs is described.
RESULTS
GFP fusion library screening for the fission yeast SG proteins.SGs are cytoplasmic aggregates that are not seen in eukaryotic cells growing under favorable conditions but are rapidly induced in response to environmental stress. Although the core structure of SGs can be biochemically purified (18), most SG components were morphologically defined using immunostaining and GFP tagging of individual proteins (19). To this end, we have taken advantage of an existing library, which covers a part of S. pombe genome (20) and further generated chromosomally tagged GFP fusion proteins, under the control of the original promoter, to determine the protein composition of the fission yeast SGs. The experiments were performed, under thermal stress (42°C for 15 min), based on the formation of granule-like structures of these GFP fusion proteins that were colocalized with the SG marker poly(A) binding protein Pabp tagged with tdTomato/mCherry. For simplicity, only green fluorescence micrographs are shown in Fig. 1. Examples of the analysis performed can be found in Fig. 2A (see also Fig. 4B). Microscopic screening of these strains detected many proteins analogous to those in mammalian SGs (Fig. 1A), including the TIA-like proteins Csx1 and Cxr1, ataxia-two homolog Ath1, the deubiquitinating enzyme Ubp3 and its cofactor G3BP homolog Nxt3 (14), the fission yeast vigilin Vgl1 (10), the AME complex (methionyl-tRNA synthetase [Rar1], glutamyl-tRNA synthetase [Gus1], and cofactor Asc1) (21) (Fig. 1B), and the early translation initiation factors eIF4G and components of the eIF3 complex (eIF3a, -b, -d, and -i) (Fig. 1C), as well as several small ribosomal proteins (Rps901, 1402, 2201, and 2501) (Fig. 1D). Intriguingly, we found that, although present in the same complex, components of the eIF3 complex and 40S ribosome were recruited to SGs differentially and that not all components of the preinitiation complex were SG proteins.
GFP fusion library screening for the fission yeast SG proteins. (A) List of fission yeast SG proteins identified in this study. (B to D) Fluorescence micrographs of strains expressing the GFP fusion proteins indicated in living cells after a temperature shift from 30 to 42°C for 15 min. Scale bars, 5 μm.
StyI-p38 as a component of SGs. (A) Merged images of fluorescence micrographs showing GFP fusion proteins of StyI (green) and Pabp-mC (red) localization in living cells grown at 30°C (no treatment), after a shift to 42°C for 15 min, or after deprivation of glucose for 1 h or an exposure to 1 M KCl for 15 min. Scale bar, 5 μm. (B) Fluorescence micrographs of the indicated strains expressing StyI-GFP protein grown at 30oC or after a shift to 42°C for 15 min. (C) Fluorescence micrographs of the indicated strains expressing Pabp-GFP protein grown at 30oC, after a shift to 42°C for 15 min, or after deprivation of glucose for 1 h or an exposure to 1 M KCl for 15 min. (D) Fluorescence micrographs showing human MAPK p38 associated with SGs in HeLa cells treated with 100 μM NaAsO2 for 1 h, but not with 1 mM thapsigargin, for 1.5 h. Cells were stained with antibody against p38α and G3BP to reveal the structure of SGs. Scale bar, 10 μm.
In addition, Puf1 and Puf4 of the fission yeast PUF family of RNA binding proteins from the GFP library were identified as the fission yeast SG proteins (Fig. 1B), which were subjected to further characterization in the experiments described below.
StyI-p38 as a component of SGs.Furthermore, the fission yeast mitogen-activated protein kinase (MAPK) StyI was also identified in this study (Fig. 2A). However, unlike the SG marker Pabp, StyI associates with SGs only under thermal stress but not with other stimuli, such as deprivation of glucose or high salt (1 M KCl). These results are in line with the notion that the constituents of SGs varied according to the type of stress. Deletion of the upstream kinase wis1 gene did not grossly affect the relocalization of StyI to SGs (Fig. 2B). It has been suggested that signaling proteins might affect the assembly of SGs. The dependency of the assembly of SGs on StyI was determined. As shown in Fig. 2C, we found that deletion of sty1 did not grossly affect the formation of SGs under the conditions tested. These results suggested that StyI was not absolutely required for the formation of SGs. It is possible that, as proposed by others (6), these proteins use SG as the signaling hub to coordinate stress response and are not involved in its assembly.
Next, we wanted to see whether the characteristic feature of StyI identified was conserved in human. Experiment was performed to determine any association of human MAPK p38 (the StyI equilibrant) with SGs. HeLa cells were stained by indirect immunofluorescence with a monoclonal antibody against p38 (Fig. 2D). As described previously in 293T cells (22), p38 was localized mainly to the nucleus with some minor presence in the cytoplasm (lower panel). Upon arsenite treatment, a chemical used to induce SG in human cells, p38 relocalized to large cytoplasmic granules (upper panel). Costaining with G3BP, a SG marker, confirmed that p38 presented in the same aggregates as G3BP after arsenite treatment. These results indicated that, as in S. pombe, p38 was a component of SGs in human cells. Like StyI, p38 responded differentially to stimuli, which was associated with SGs with the treatment of arsenite but not with the stimuli of thapsigargin (middle panel). Together, these results suggested that the characteristic feature of the fission yeast MAPK StyI with SG was conserved in human p38.
Fission yeast PUF family of RNA binding proteins.The identification of Puf1 and Puf4 as SG proteins has drawn our attention to the possibility of the involvement of the PUF family of RNA binding proteins in SGs. The PUF protein family is widespread among species, with members found in many eukaryotes, including humans, plants, animals, and unicellular and metazoan organisms (16). There are six (referred to as Puf1, Puf2, Puf3, Puf4, Puf5, and Puf6) plus two meiotic specific proteins (Mpf1 and Mpf2) encoded in the S. pombe genome, which contain six to eight copies of the characteristic PUF repeat sequences (Fig. 3A). Little is known about the physiological functions of these proteins in S. pombe. As described above, we were interested in the possibility of the involvement of the other PUFs in SGs. As a first step toward determining the biological functions of the PUF family proteins in S. pombe, targeted recombination was used to generate GFP fusion proteins in its normal chromosomal context under the control of the endogenous promoter, expressing the full-length protein with a C-terminal GFP tag. Western immunoblotting demonstrated the expression of all six GFP-tagged PUF proteins with Puf1 of the highest and Puf5 of the lowest expression levels (Fig. 3B). Despite these results, however, Puf5 failed to generate detectable signals under microscopic examination, possibly due to its low abundance, and was not pursued further. As shown in Fig. 3C, in accordance with the proposed role in mRNA expression, PUF proteins presented predominately in cytoplasm of the cells with the exception of Puf6, which was found mainly in the nonchromosomal hemisphere. The involvements of these proteins in SGs were determined in the following experiments described.
Fission yeast PUF family of RNA binding proteins. (A) Schematic representation of the domain structures of the fission yeast PUF family proteins. (B) Whole-cell protein extracts of the strains indicated were prepared by alkaline extraction, followed by trichloroacetic acid precipitation. The extracts were separated by SDS-PAGE and subjected to immunoblotting using anti-GFP antibodies to reveal PUF proteins. Antibody against α-tubulin (tub) was used as the control. The relative level of PUF protein is indicated beneath each lane (average from two independent experiments). (C) Fluorescence micrographs of strains expressing GFP fusion PUF proteins with Hoechst 33342 staining to reveal nucleus. Scale bar, 5 μm. (D) Cellular localization of the fission yeast PUF family proteins.
PUFs are SG components.As a first step toward determining the involvement of PUF proteins in the fission yeast SGs, we examined the localization of Puf1, -2, -3, and -4 proteins under various types of stress, which would induce SG formation (Fig. 4A). As in the scenario of Puf1 and Puf4, thermal stress caused the relocalization of Puf2 and -3 to distinct cytoplasmic granule-like structures. Similarly, after the exposure of cells to 1 M KCl or with glucose starvation for 1 h, a rapid relocalization of Puf1, -2, -3, and -4 proteins was observed. These were different from StyI, which only responded to thermal stress. In addition, we observed that Pabp colocalized almost always with GFP fusions of Puf1, -2, -3, and -4 proteins under thermal stress (Fig. 4B). This colocalization indicates that these proteins are components of a single granule. The essential roles of these proteins were also explored. We found that the deletion of individual puf1, -2, -3, and -4 genes specifically affected the formation of SGs after glucose starvation but not under thermal stress (Fig. 4C). These results suggested that not only did the components of SGs vary according to the type of stress, the underlying molecular mechanisms for their assembly could be different, too. No additional defects were observed when simultaneously deleted all four genes, suggesting that these proteins worked in a similar fashion. Together, these results identified Puf1, -2, -3, and -4 as the fission yeast SG proteins that were required for the assembly of SGs after deprivation of glucose. In line with these results, as shown in Fig. 4D, puf mutants were defective in recovery from stationary phase after glucose depletion with a much longer time to reenter cell cycle.
Puf1, -2, -3, and -4 were components of fission yeast SGs. (A) Strains expressing GFP fusion of PUF proteins were visualized after the growth to mid-logarithmic phase at 30°C (no treatment), after a shift to 42°C for 15 min, or after deprivation of glucose for 1 h or an exposure to 1 M KCl for 15 min. Scale bar, 5 μm. (B) Merged images of fluorescence micrographs showing GFP fusion PUF proteins (green) and Pabp-mC/tdT (red) after a 15-min incubation at 42oC. (C) Fluorescence micrographs of the indicated strains expressing Pabp-GFP protein grown at 30°C (no treatment), after a shift to 42°C for 15 min, or after deprivation of glucose for 1 h. The percentages of 200 cells positive for SGs after deprivation of glucose are indicated for the strains tested. Data from two independent experiments are expressed as means ± the standard deviations (SD). Asterisks indicate the statistical significance (P < 0.05) versus puf+. (D) Growth characteristics of the strains indicated at 30°C (doubling time) or refed from stationary phase after glucose depletion to reenter the cell cycle.
PUFs function in SGs.Recent studies have shown that interactions between low-complexity and intrinsically disordered protein regions can facilitate proteins concentrated into discrete subcellular domains such as SGs (23). Keeping with the function in SG assembly (Fig. 4C), several low-complexity and intrinsically disordered protein regions were identified within Puf1, -2, -3, and -4 proteins sequences (Fig. 5A). In line with these results, we found that overexpression of Puf1, -2, -3, and -4 from the multicopy pREP1 plasmid, under the control of the thiamine-repressive nmt1 promoter, led to the formation of multiple enlarged foci (Fig. 5B). These results were not observed with overexpression of the other SG proteins identified (14), suggesting that Puf1, -2, -3, and -4, but not other SG proteins, tended to form aggregates, which might contribute to the formation of SGs. In support of this idea, overexpression of Puf1, -2, 3, and -4 induced the aggregation of Pabp, the signature constituent of SGs (Fig. 5B).
Involvement of PUFs in SG formation. (A) Schematic representation of the low-complexity and intrinsically disordered protein regions in the fission yeast PUFs. (B and C) Fluorescence micrographs of strains expressing Pabp-mCherry also overexpressing (OP) full-length (B) or truncated (C) PUFs from the multicopy pREP1 plasmid under the control of the thiamine-repressive nmt1 promoter. Scale bar, 5 μm.
The involvement of the low-complexity and intrinsically disordered protein regions in protein aggregation was also tested. As shown in Fig. 5C, truncation of the low-complexity and intrinsically disordered protein regions affected the aggregation of Puf1, -2, -3, and -4 proteins to form granule-like structures. In line with the function in SG assembly, the truncated Puf2 protein failed to induce aggregation of Pabp.
Intriguingly, a distinct structure, which colocalized well with Pabp, was observed with the truncation of Puf1, -3, and -4 proteins. These results suggested that these PUFs might have an additional function, which was not shared by Puf2.
PUFs function in PBs.Recently, PUM2, a human PUF, was identified in the PB proteome (24). Intriguingly, we found that Puf1, -2, -3, and -4, when overexpressed, also affected the structure of PBs, with multiple enlarged foci revealed by Dcp2-tdT (Fig. 6A). These results suggested a function of these PUFs in PBs. Unlike overexpression from the multicopy pREP1 plasmid (Fig. 5B), we noticed that, when overexpressing from the single genomic integrated nmt1 promoter, these PUF aggregates affected the localization of Dcp2 but not that of Pabp. Examples of overexpression of Puf1 from genomic integrated nmt1 promoter are shown in Fig. 6B. These results suggested that, in comparison to PBs, relatively high-level local concentrations of PUFs were required to induce SG formation.
Involvement of PUFs in PB function. (A) Fluorescence micrographs of strains expressing Dcp2-tdTomato also overexpressing (OP) PUFs from the multicopy pREP1 plasmid under the control of the thiamine-repressive nmt1 promoter. Scale bar, 5 μm. (B) Fluorescence micrographs of strains expressing Dcp2-tdT/Pabp-mC also overexpressing (OP) Puf1 from the genomic integrated nmt1 promoter. For simplicity, only fluorescence micrographs of proteins indicated are shown. (C) Merged images of fluorescence micrographs showing Pabp-GFP (green) and Dcp2-tdTomato (red) proteins after deprivation of glucose. (D) Merged images of fluorescence micrographs showing GFP fusion PUFs (green) and Dcp2-tdTomato (red) proteins 15 min after deprivation of glucose. (E) Fluorescence micrographs of the indicated strains expressing Pabp-GFP protein grown at 30°C (control) or after deprivation of glucose for 1 h. The percentages out of 200 cells positive for SGs after deprivation of glucose are indicated for the strains tested. Data from two independent experiments are expressed as means ± the SD. Asterisks indicated statistical significance (P < 0.05) versus the wild type. (F) Fluorescence micrographs of strains expressing Pabp-mC also overexpressing (OP) proteins indicated from the nmt1 promoter. (G) Fluorescence micrographs of the indicated strains expressing Dcp2-GFP protein grown at 30°C (control) or after deprivation of glucose for 15 min.
These results also suggested a close connection between SGs and PBs, which was further explored by studying the kinetics of their formation after glucose starvation, the stimuli for both structures. Unlike in S. cerevisiae and similar to the situation in mammalian cells, small, microscopically visible PBs (Dcp2 foci) are constitutive in S. pombe (4). As the previously described characteristic feature of PB, Dcp2 foci were enhanced after deprivation of glucose for 15 min (Fig. 6C), with an increase in size and intensity, which were colocalized with Puf1, -2, and -3 to a certain degree (Fig. 6D). We have previously demonstrated that, unlike other components of PBs, the recruitment of the deadenylase Ccr4 proteins to PBs only occurred upon stress (5). It has been demonstrated that Ccr4 interacted with Puf3 in S. pombe (25). These results suggested that, similar to Ccr4, Puf1, -2, and -3 might be involved in some aspect of PB function and were recruited to these foci after glucose starvation.
PBs promote SGs assembly in S. pombe.Keeping with the connection of SGs and PBs, we found that, 30 min after glucose starvation, the formation of the Pabp foci occurred, and they colocalized well with the Dcp2 foci (Fig. 6C). These results suggested that, as in S. cerevisiae (9), PBs and SGs somewhat overlapped in S. pombe. The proceeded and overlapped structure of SGs with PBs upon glucose starvation led us to speculate that PB might have a role in SG assembly. In support of this hypothesis, as shown in Fig. 6E, the formation of SGs was compromised in mutants with pdc1 and pdc2 deleted, which were required and enhanced the formation of PBs when overexpressed (4, 5). However, Pdc1 and Pdc2 overexpression did not affect the localization of Pabp (Fig. 6F). These results suggest that PB is not the sole factor and required additional proteins such as PUFs to induce SG formation. The dependency of the assembly of PBs on PUFs was also demonstrated. As shown in Fig. 6G, we found that simultaneous deletion of puf1, -2, -3, and -4 genes did not grossly affect the formation of PBs before and after glucose starvation. Together, these results suggested that, as in S. cerevisiae (9) and human cells (26), PBs promote SG assembly in S. pombe. PUFs are involved in both structures and may provide the link for SGs to PBs.
Overexpression of Puf2 represses translation.The results described above suggested a function of PUFs in PBs. The best-characterized function of PUFs is as posttranscriptional repressors (16). It is possible that PUFs function as regulators of translational repression in PBs. Intriguingly, while in S. cerevisiae, the PB protein Pat1 is a regulator of both mRNA decay and translational repression, the roles of human Pat1b and the Pdc2/Pat1 in S. pombe appear to be restricted to mRNA degradation (5). It is tempting to speculate that, in other organisms, PUFs may function to replace the role of Pat1 in translational repression. In support of this idea, overexpression of Puf1, -2, -3, and -4 led to growth arrest in S. pombe cells (Fig. 7A). Furthermore, we found that truncation of the low-complexity and intrinsically disordered protein regions relieved the growth inhibitory effect of Puf2 but not that of the Puf1, -3, and -4 proteins (Fig. 7B). As described above, Puf1, -3, and -4 might possess functions not shared by Puf2. These proteins could work through additional mechanisms for translational repression that required more study and was not pursued further here.
Puf2 repressed translation for SG formation. (A to C) Tenfold serial dilutions of strains transformed with pREP1 (control) or pREP1 encoding the wild-type or truncated PUF protein indicated were spotted onto EMM-agar plates with (OFF) or without (ON) thiamine and photographed after 2 days incubation at 30°C. (D) Native cytosolic extracts from CHX-treated wild-type and Puf2 mutant (1-592) overexpression (OP) cells grown at 30°C were separated on 7 to 47% sucrose gradients. The distribution of the rRNAs is shown by the absorption profiles at 260 nm. (E) Merged images of fluorescence micrographs showing Pabp-GFP (green) and Dcp2-tdTomato (red) proteins in cells overexpressed Puf2 (1-592) from the multicopy pREP1 plasmid under the control of the thiamine-repressive nmt1 promoter. (F) Whole-cell protein extracts from indicated strains treated with puromycin for 2 h were prepared by alkaline extraction, followed by trichloroacetic acid precipitation. As a positive control, the cells were pretreated with cycloheximide (CHX) at 40 μg/ml for 2 h. The extracts were separated by SDS-PAGE and subjected to immunoblotting using antipuromycin antibody to reveal the level of incorporation. The relative levels of puromycin incorporation are indicated beneath each lane (average from three independent experiments). (G) Coimmunoprecipitation was performed with extracts prepared from eIF4G-GFP tagged strains expressing Puf2 tagged with flag. GFP-Trap affinity resin was used to pull down GFP proteins. Immunoprecipitates (IP) were then analyzed by Western immunoblotting with antibodies against GFP and flag. An asterisk indicates the nonspecific band cross-reacted with GFP antibody. (H) Coimmunoprecipitation was performed with extracts prepared from Pdc1-GFP-tagged strains expressing Puf2 tagged with flag. GFP-Trap affinity resin was used to pull down GFP proteins. Immunoprecipitates (IP) were then analyzed by Western immunoblotting with antibodies against GFP and flag. An asterisk indicates the nonspecific band cross-reacted with GFP antibody.
It has been suggested that PUFs bind to specific recognition sequence in the 3′ untranslated region of mRNA to control the translation of the transcript. The results described above suggested that the growth inhibitory effect of Puf2 could be independent of its RNA-binding activity. In line with this idea, overexpression of Puf2 mutant protein (i.e., 1-591), which lacked both the RRM and PUF RNA-binding domains, led to growth arrest in S. pombe cells (Fig. 7C). To show that this is due to translational repression, we investigated the polysome profile of puf2 mutant (Fig. 7D). When the Puf2-mutant protein overexpressed, a marked reduction of polysomes assembled and increase in the relative proportion of 40S and 60S ribosomal peaks were observed. We further measured the rates of global protein synthesis in the presence of puromycin, which became incorporated into nascent polypeptide chains by ribosomal catalysis, to label newly synthesized proteins (21). As shown in Fig. 7F, the puromycin-incorporated proteins were significantly reduced when overexpressing the full length and the N-terminal region (1-592 amino acid) of the Puf2 protein (lanes 3 and 4 compared to lane 2). Overexpression of the C-terminal region (amino acids 712 to 1065) of the Puf2 protein did not grossly affect global protein synthesis, as measured by puromycin incorporation (lane 5). As a positive control, pretreatment of cells with cycloheximide (CHX), which inhibited translation elongation by trapping the mRNA in polysomes, led to the decrease of labeled proteins (lane 1). Together, these results suggested that Puf2 might repress translation for PB and SG formation, which was independent of its RNA binding activity. In line of this idea, overexpression of Puf2 mutant protein (1-591) induced aggregation of Dcp2-tdT and Pabp-GFP (Fig. 7E). Keeping with the function of PB in SG formation, these Puf2-induced Pabp-GFP aggregates (SGs) were associated with Dcp2-tdT (PBs).
Puf2 links SGs to PBs.The results described above suggested a function of Puf2 linking SGs to PBs. It has been demonstrated that, in S. cerevisiae, SG protein eIF4G can be detected to accumulate to some degree in PBs under glucose deprivation (27). We speculate that Puf2 might be interacted with eIF4G to bridge the interaction between SGs and PBs. In support of this idea, as shown in Fig. 7G and H, we found that Puf2 specifically interacted with eIF4G and Pdc1, as demonstrated by coimmunoprecipitation, but not with the GFP control. These results suggested that Puf2 might interact with multiple factors in SGs and PBs to bridge the interaction between these two structures.
Meiotic PUF granules associated with PBs.The results described above suggested a close interaction between PUFs and PBs. During characterization of the fission yeast PUFs, we found that Mpf2, the meiotic PUFs, also formed cytoplasmic granule-like structures in meiotic cells, which were not seen in interphase cells (Fig. 8A and B; the absence of Mpf2 granule-like structures in interphase cells is indicated by arrowheads). We referred to these granules as meiotic PUF granules (MPGs). Experiments were performed to determine whether these granules might be function related to SGs or PBs. As the previously described characteristic feature of SGs, nitrogen starvation did not induce SG formation (10), and Pabp foci were not present in meiotic cells (Fig. 8B). In contrast, Dcp2 foci (PBs) presented throughout meiosis, which colocalized with MPGs to certain degree (Fig. 8A), suggesting a meiotic function of these two types of granules. The dependency of the assembly of MPGs on PB was also determined. As shown in Fig. 8C, we found that the MPGs could still form in pdc1 and pdc2 mutants. Deletion of mpf1 and mpf2 did not grossly affect the structure of Dcp2 (Fig. 8D). These results suggested that these two structures could form independently. Furthermore, we found that the mpf1Δ mpf2Δ mutant could still carry out meiosis/sporulation with reduced spore viability. These results suggested that Mpf1 and Mpf2 were not absolutely required for meiosis.
Meiotic PUF granules associated with PBs. (A and B) Fluorescence micrographs of strains expressing Mpf2-GFP (green) also overexpressing Dcp2-tdTomato (red) proteins (A) and Pabp-mCherry (red) proteins (B) in meiotic cells. Scale bar, 5 μm. Arrowheads indicate the absence of Mpf2-GFP granules in interphase cells. (C and D) Fluorescence micrographs of meiotic cells expressing Mpf2-GFP (C) or Dcp2-tdT (D) proteins in the strains indicated. (E) Micrographs of meiotic cells of the strains indicated. Arrowheads indicate aberrant spores. The viabilities of 200 randomly selected spores by tetra dissector from two independent experiments of the strains indicated are expressed as means ± the SD. Asterisks indicate the statistical significance (P < 0.05) versus the wild type.
DISCUSSION
The characteristic feature of the fission yeast SGs.To gain more insight into the protein composition and structure organization of the fission yeast SGs, in this study, we described a GFP fusion library screening for the fission yeast SG proteins (Fig. 1). Several interesting aspects were observed. We found that they contained many proteins analogous to those in mammalian SGs, including components of the 48S preinitiation complex (small ribosomal subunits as well as the early translation initiation factors) and specific RNA-binding proteins. Intriguingly, we found that not all components of the preinitiation complex were in SGs. These results suggested that, despite present in the same complex, they were recruited to SGs differentially. In line with this result, in an analysis of the disordered region within human 40S ribosomal subunit proteins (28), several proteins were found to contain regions that might facilitate them to concentrate into SGs. Alternatively, the complex might be processed before recruited to SGs. However, there is no evidence to suggest this.
In addition to their role as mRNA triage centers, it has been suggested that SGs might function as the signaling hub (6) to coordinate stress response. In support of this idea, we found that the fission yeast MAPK StyI and human p38 associated with SGs (Fig. 2). Furthermore, we found that, depending on the type of stress, the components of SGs varied and the underling molecular mechanism required for their formation could be different, too. Members of the fission yeast PUF family of RNA-binding proteins including Puf1, -2, -3, and -4 were identified as the fission yeast SG proteins that were required specifically for SG formation upon glucose starvation (Fig. 4). No additional defects were observed in a four-way mutant, suggesting that they function in a similar fashion. Although Puf1, -2, -3, and -4 associated with SGs under various types of stress, it is not clear whether these proteins play similar roles with other stimuli. Given the highly redundant nature of proteins involved in assembly of these macro ribonucleoproteins aggregates, it is likely that, along with other stimuli, additional factors are required and can functionally substitute PUFs.
PUFs function in SGs and PBs.In line with a function in SG formation, Puf1, -2, -3, and -4 contain several low-complexity and intrinsically disordered protein regions with a tendency to form aggregates (Fig. 5B) that were not shared with the other SG proteins (14). It has been suggested that interactions between low-complexity and intrinsically disordered protein regions can facilitate proteins concentrated into SGs (23). In keeping with this idea, when overexpressed, Puf1, -2, -3, and -4 induced aggregation of Pabp, the signature constituent of SGs (Fig. 5B). Truncation of the low-complexity and intrinsically disordered protein regions affected the aggregation of Puf1, -2, -3, and -4 proteins (Fig. 5C).
Our data also suggested a role of PUFs in PBs, and components of PBs could be divided into two classes in S. pombe. The core components of PBs including Exo2, Dcp1-Dcp2, Pdc1, Edc3, Ste13, Pdc2, and Lsm1 are constitutively visible in PBs (4, 5). In contrast, components such as Ccr4 (5) and PUFs demonstrated in this study associated with PB foci only during stress response (Fig. 9B). PUFs might function as the regulators of translational repression. In line with this idea, Puf1, -2, and -3 were recruited to PBs after glucose starvation (Fig. 6D). When overexpressed, Puf1, -2, -3, and -4 enhanced the formation of PBs (Fig. 6A) and led to growth arrest in S. pombe cells (Fig. 7A). An RNA binding-independent translational repression activity of Puf2 was identified (Fig. 7D). The interaction with eIF4G (Fig. 7G) suggested that Puf2 might interact with the preinitiation complex to repress translation upon stress (Fig. 9C). Translation repression increased the pools of nontranslating RNAs, which were partitioned according to their characteristic features, such as poly(A) length, to associated with PB or SG proteins (29). The low-complexity and intrinsically disordered protein regions of Puf2 helped to facilitate these proteins concentrated into discrete subcellular domains.
Puf2 functions in SGs and PBs. (A) Schematic representation of the domain structures and functions of Puf2 proteins. (B) Characteristic features of PB components. (C) Model summarizing the involvement of Puf2 in SG and PB formation. Upon glucose starvation, Puf2 interacts with the preinitiation complex to repress translation. Translation repression increases the pools of nontranslating RNAs, which are partitioned according to their characteristic features, such as poly(A) length to associated with PB or SG proteins. The low-complexity and intrinsically disordered protein regions of Puf2 helps to facilitate these proteins concentrated into discrete subcellular domains.
The involvement of Puf2 in SGs and PBs formation suggested that these two structures interacted. Although a connection between SGs and PBs has been reported (9, 26), how they interacted remained to be determined. The observation that, upon glucose starvation, SG aggregation proceeded and overlapped with the structure of PBs (Fig. 6C) led us to suggest that, although these structures could form independently, PB might function as the scaffold to promote the formation of SGs (Fig. 9C). In support of this idea, SG formation was compromised in pdc1 and pdc2 deletion mutants (Fig. 6E), since these genes are required for PB formation (4, 5). Additional evidence is needed to consolidate this hypothesis. Nevertheless, our results suggested a close interaction of these two structures, and PUFs might provide the link between them.
Furthermore, we found that the interaction between PUFs and PBs could be further extended to meiotic cells. MPGs, a novel meiotically specific type of PUF granules that are associated with PBs, were identified (Fig. 8). These results suggested a meiotic function of these two structures. In line with this idea, the fission yeast PB proteins Ste13 and DEXH/D-box RNA helicase 1 (Dhh1; RCK/p54 in mammals) were identified in a screen for mutants defective in meiosis (30). Further characterization of MPGs would provide novel insights into the regulation of meiosis. Although we did not address this question, the involvement of PUFs in SG and PB functions described here added significantly to the versatility and complexity of the pathway and the mechanism in which members of PUF family of RNA-binding proteins participated.
MATERIALS AND METHODS
Fission yeast strains, plasmids, and methods.Conditions for growth, maintenance, and genetic manipulation of fission yeast were as described previously (10). A complete list of the strains used in this study is given in Table 1. One-step gene disruption or modification via homologous recombination was performed following PCR-mediated generation of ura4+ or KanMX selectable cassettes flanked by 80-bp segments from appropriate regions of the genes of interest using oligonucleotides described in Table 2. Except where otherwise stated, strains were grown at 30°C in yeast extract or Edinburgh minimal medium (EMM) with appropriate supplements. Where necessary, gene expression from the nmt1 promoter was repressed by the addition of 15 μM thiamine to the culture medium.
Fifty-nine S. pombe strains used in this studya
Oligonucleotides used in this study
Plasmids for the expression or the truncation of PUFs were constructed by PCR amplification of the corresponding open reading frame of S. pombe cDNA, using the primer pairs described in Table 2, and subsequently ligated into plasmids derived from pREP1 with a GFP or glutathione S-transferase (GST) tag inserted at the 3′ end of the multiple cloning sites.
GFP fusion library screening.To avoid autofluorescence from rich medium, Pabp-tdT/mC strains expressing the GFP fusion of candidate proteins were grown in EMM to mid-logarithmic phase at 30°C. Then, 10-ml exponentially growing cultures were concentrated to 200 μl in MCC tubes to facilitate microscopy observation before putting them on a 42°C heat block for 15 min. Cell suspensions (1 μl) were applied onto poly-l-lysine-coated slides covered with cover slides for live cell fluorescence microscopy at room temperature. The analysis was based on the formation of granule-like structures of these GFP fusion proteins that were colocalized with the SG marker poly(A) binding protein Pabp.
Cell recovery upon glucose deprivation.Yeast strains were grown in YES medium at 30°C overnight to stationary phase after being glucose deprived (>5 × 107 cells/ml). The cells were collected by centrifugation and cultured in fresh YES medium at a 10-fold dilution to determine the time required to double their cell numbers at 30°C.
Antibodies and immunoprecipitation.Yeast extracts prepared by alkaline extraction as described previously (10) were resolved by SDS-PAGE before being subjected to Western immunoblotting with anti-GFP antibody (Abcam, Cambridge, UK) to reveal GFP fusion proteins. Anti-flag M2 antibody was from Sigma-Aldrich (St. Louis, MO). Antibody to α-tubulin (Sigma-Aldrich) was used as a control.
For immunoprecipitation, 4 × 108 cells were lysed in 300 μl of NP-40 buffer (6 mM Na2HPO4, 4 mM NaH2PO4, 1% NP-40, 150 mM NaCl, 2 mM EDTA, 50 mM NaF, 0.1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, Complete protease inhibitor cocktail) by vortexing with acid-washed glass beads. The lysate was clarified by centrifugation and GFP fusion proteins were retrieved using GFP-Trap-coupled agarose beads (Chromotek, Martinsried, Germany) according to the manufacturer’s instructions. Next, 5% of the input was loaded onto the gel.
Microscopy.Visualization of tdTomato-, mCherry-, and GFP-tagged proteins in living cells was performed at room temperature. In some experiments, living cells growing in EMM were stained by the addition of 5 μg/ml bis-benzamide (Hoechst 33342; Sigma) before examination by fluorescence microscopy. Images acquired from a Leica DM RA2 microscope equipped with a Leica DC 350F camera were assembled using Adobe PhotoShop.
Cell culture and immunofluorescence.HeLa cells were cultured in Dulbecco modified Eagle medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal calf serum, 100 IU/ml penicillin, and 0.1 mg/ml streptomycin and maintained at 37°C in a humidified incubator containing 5% CO2 in air. To induce SG formation, HeLa cells were treated with 100 μM NaAsO2 for 1 h or 1 mM thapsigargin for 1.5 h. For immunofluorescence, HeLa cells grown on coverslips were fixed with 4% formaldehyde and stained with rabbit anti-p38α monoclonal antibody (Cell Signaling Technology) diluted 1:150 in phosphate-buffered saline (PBS) containing 1% bovine serum albumin, 0.05% saponin, and 0.05% sodium azide. Dylight 594-conjugated secondary antibody (Jackson ImmunoResearch Laboratories) was used in a dilution of 1:200. Cells were costained with mouse anti-G3BP monoclonal antibodies (Abcam, Cambridge, UK) at a dilution of 1:150 and Dylight 488-conjugated secondary antibody (Jackson ImmunoResearch Laboratories) to reveal the structure of SGs. Cells were mounted using 50% glycerol in a high-quality antifade medium (Abcam). Images acquired from a Leica DM RA2 microscope equipped with a Leica DC 350F camera were assembled using Adobe Photoshop.
Polysome profile analysis.Polysomes were obtained as previously described (5). Cycloheximide (100 μg/ml) was added to 50-ml portions of cultures at an optical density at 600 nm of 0.5 grown at 30°C. The cultures were harvested and processed for polysome profiling on 7 to 47% sucrose gradients. Profiles were obtained by online measurements of the A260.
Puromycin incorporation assay.Yeast cells were treated or left untreated with cycloheximide (40 μg/ml) for 2 h. Then, 20 μg/ml puromycin (InvivoGen, San Diego, CA) was added 2 h before harvest. The levels of puromycin incorporation in each cell lysates were determined by immunoblotting with puromycin monoclonal antibodies (Merck Millipore, Billerica, MA).
ACKNOWLEDGMENTS
We thank Takashi Toda, Yasushi Hiraoka, and the Yeast Genetic Resource Center Japan for yeast strains and Ching-Ting Huang for technical support.
This study was supported by the National Health Research Institutes of Taiwan (MG-108-PP-08).
FOOTNOTES
- Received 18 November 2019.
- Returned for modification 23 December 2019.
- Accepted 12 February 2020.
- Accepted manuscript posted online 18 February 2020.
- Copyright © 2020 American Society for Microbiology.
