The polysome-associated proteins Scp160 and Bfr1 prevent P body formation under normal growth conditions

The protein content in a cell at any given time is regulated by transcription, translation and protein degradation. Transcription largely determines the number of mRNA molecules, whereas the balance of mRNA translation and protein degradation is important for the regulation of protein homeostasis. Upon encountering stress, a cell needs to quickly alter its proteome, thus, as a type of primary response to stress, translation often becomes attenuated. Consequently, mRNAs that are no longer engaged in translation are frequently stored in processing bodies (P bodies) (Parker and Sheth, 2007). P bodies consist of the decapping enzymes Dcp1 and Dcp2, the helicase Dhh1, activators of decapping – such as Pat1, Scd6, Edc3 and the Lsm1–7 complex – and the 5′-3′ exonuclease Xrn1 (Parker and Sheth, 2007). In addition to this basic assortment of P body components, an ever-growing list of proteins associated with P bodies has been reported (Jain and Parker, 2013; Mitchell et al., 2013; Parker and Sheth, 2007).

To date, P bodies have been reported to be involved in mRNA decapping, nonsense-mediated decay, translational repression, mRNA storage and other RNA related functions (Parker and Sheth, 2007). The transcripts that are stored in P bodies are in a dynamic exchange with the pool of mRNAs undergoing translation and can either be degraded or be bound by ribosomes for translation.

P bodies are not the only ribonucleoprotein (RNP) structures that are formed upon stress. Numerous stress conditions also promote the formation of stress granules. Stress granules are mainly composed of stalled translation initiation components and are sites of mRNA storage. The analysis of the different RNP structures is complicated by the fact that numerous factors are found in both types of granules, such as the exonuclease Xrn1 (Buchan and Parker, 2009). In addition, mRNA can traverse from P bodies into stress granules and vice versa (Buchan et al., 2008). Moreover, P bodies and stress granules are found in close contact to each other, which might facilitate the transfer of components and mRNA.

In yeast, 5′-to-3′ decay of mRNA appears to be the major direction of mRNA degradation (Muhlrad et al., 1995; Shoemaker and Green, 2012). Once the cell encounters stress, P bodies become visible by microscopy, which is thought to occur through the aggregation of P body components that are, presumably, bound to mRNA. These aggregates are, nevertheless, highly dynamic and can dissolve either upon adaptation of the cell to the stressor or when the stress ceases to exist (Brengues et al., 2005; Teixeira and Parker, 2007). Depending upon the stress that is encountered, the size and number of P bodies varies. For example, glucose starvation induces a few relatively large foci, whereas high salt and high Ca²⁺ cause the formation of smaller but more numerous P bodies (Brengues et al., 2005; Kilchert et al., 2010). The reason for the different responses is not clear. It has to be noted, however, that glucose starvation can only be remedied by the addition of a good carbon source, whereas cells can adapt to stresses such as high salt (Ashe et al., 2000; Brengues et al., 2005). As a consequence, P bodies that form under such stress disassemble after adaptation (Kilchert et al., 2010).

P bodies are found in close proximity to, and might even be physically linked to, the endoplasmic reticulum (ER) (Kilchert et al., 2010). To determine whether P bodies are physically linked to the ER, we aimed to identify proteins that bind to P body components. The peripheral ER proteins Scp160 and Bfr1 were identified as interacting proteins of P bodies. The association of Scp160 and Bfr1 with polysomes prompted us to test whether Scp160 and Bfr1 might play a role in P body formation. Our results suggest that Bfr1 and Scp160 act independently as negative regulators of P body formation at the ER. Moreover, our data imply that Scp160 is required for the correct assembly of P bodies.

We have shown previously that P bodies are localized in close proximity to ER membranes (Kilchert et al., 2010). In order to identify potential ER-localized interaction partners of P body components, we employed tandem affinity purification after crosslinking with the P body components Dcp2 and Scd6, which are part of the 5′UTR- and the 3′UTR-associated complexes of P bodies, respectively. We chose a crosslinking approach because P bodies are highly dynamic, and their interaction with the ER could be transient (Teixeira and Parker, 2007). We also opted for a stringent method and, therefore, selected the HBH tag for the purification scheme. The HBH tag consists of a biotinylation sequence, which is flanked by two His₆ peptides, allowing tandem purification under denaturing conditions (Tagwerker et al., 2006).

The temperature-sensitive arf1-11 mutation in the small GTPase Arf1p causes the formation of numerous P bodies at the restrictive temperature (Kilchert et al., 2010). Cultures of arf1-11 mutants expressing chromosomally tagged Dcp2–HBH or Scd6–HBH and wild-type cells were shifted to 37°C for 1 h to induce P bodies. Protein complexes were crosslinked in vivo, extracted from membrane pellets that had been enriched for ER membranes, and purified over Ni-NTA and streptavidin beads under denaturing conditions. The proteins were digested, and the peptides analyzed by mass spectrometry. This approach was successful because we could identify P body components in our purification scheme (Table 1). In addition, proteins were identified that were neither present in the untagged wild-type control nor in other HBH purifications for unrelated baits that have been performed by our group (Ritz et al., 2014). Those proteins present only in the Dcp2–HBH or Scd6–HBH sample, or in both samples, were considered for further analysis (Table 1).

Three of the proteins identified in the tandem affinity purifications were ER-associated proteins – Bfr1, Scp160 and Cdc48. Cdc48 is a AAA-ATPase that is involved in homotypic membrane fusion during karyogamy, and in the extraction and retro-translocation of polypeptides from the ER that are destined for degradation by the proteasome (Hitchcock et al., 2001; Latterich et al., 1995; Schuberth and Buchberger, 2005). The function of Cdc48 in ER-associated degradation might mean that it also transmits a signal to induce P body formation under ER stress conditions. We decided, however, to focus on the two other proteins Scp160 and Bfr1. Scp160 is an mRNA-binding protein that, together with Bfr1, associates with actively translating polysomes at the ER (Lang et al., 2001). Loss of Scp160 causes mislocalization of the asymmetrically localized ASH1 mRNA (Irie et al., 2002; Trautwein et al., 2004) and SRO7 mRNA after pheromone treatment (Gelin-Licht et al., 2012). Bfr1 has been identified as a multicopy suppressor of brefeldinA-induced lethality (Jackson and Képès, 1994). However, the molecular function of Bfr1 is still unknown.

First, we established whether Scp160 and Bfr1 are still associated with the ER at 37°C, the temperature at which we performed the crosslink. To this end, we appended both Scp160 and Bfr1 chromosomally with a GFP tag in strains that expressed the ER marker Sec63 that had been tagged with red fluorescent protein (RFP). Both proteins remained associated with the ER, irrespective of the temperature (23°C or 37°C) or the strain background (wild-type or arf1-11) (Fig. 1).

Because we used a crosslinking approach to detect proteins that interact with P body components, we wanted to ensure that Scp160 and Bfr1 were not just localized in the vicinity of P bodies and therefore caught in the complex, and that Scp160 and Bfr1 were also able to interact with P body components in the absence of crosslinking agents. 3× FLAG tagged Scp160 co-immunoprecipitated Dcp2 that had been tagged with nine repeats of a Myc epitope (9Myc), irrespective of whether the cells were stressed by glucose starvation, high intracellular Ca²⁺ or unstressed (Fig. 2A,B). The association of Bfr1 with Dcp2 was detected to a lower extent, and the levels were more variable, indicating that Dcp2 might interact primarily with Scp160 rather than with Bfr1. Given that Scp160 is an RNA-binding protein and interacts with the machinery that stores and decays mRNAs, we next tested whether the interaction between Scp160 and Dcp2 was dependent on mRNA. Indeed, treatment of yeast lysates with RNase strongly reduced the amount of co-immunoprecipitated Dcp2 (Fig. 2C).

To explore the possibility that other P body components could also interact with Scp160 and/or Bfr1, we repeated the immunoprecipitations and probed for Pat1 and Edc3, both of which had been tagged with six repeats of the HA epitope (Pat1–6HA and Edc3–6HA, respectively), without stress or under glucose starvation (Fig. 2D). Similar to Dcp2, Pat1 and Edc3 were part of a complex that was associated with, at least, Scp160–FLAG. In all cases, the stress seemed to slightly, but consistently, increase the interaction of Scp160 with these P body components. By contrast, Scd6–6HA was not co-immunoprecipitated with Scp160–FLAG (Fig. 2D). The ability of Scp160 to interact with P body components appeared to be independent of stress, indicating that induction of stress is not a prerequisite for the interaction of P body components with Scp160. Because Dcp2 and Pat1 play a more prominent role in P body assembly (Teixeira and Parker, 2007), our data suggest that at least partially assembled P bodies could be present on actively translating ribosomes. We conclude that Scp160 is a newly identified mRNA-dependent protein that interacts with P body components at the ER.

P bodies localize in close proximity to the ER (Kilchert et al., 2010). Therefore, we asked whether Scp160 and Bfr1 are required for P body localization at the ER. Dcp2–9Myc-positive structures were detected by immunoelectron microscopy as being in close proximity to the ER in both the Δbfr1 and Δscp160 cells, or in the wild type, when shifted to 37°C (Fig. 3A). The distribution of the Dcp2–9Myc signal, which was suggestive of a ring-like globular structure, was similar in both strains. We then used a biochemical assay to investigate further the results of the ultrastructural analysis. We have shown previously that Dcp2–9Myc co-migrates with ER membranes in a sucrose gradient (Kilchert et al., 2010); here, the association of Dcp2 with the ER was not drastically altered in either of the deletion strains (Fig. 3B) – similar amounts of Dcp2–9Myc floated at the 0% to 40% sucrose interface. Our data demonstrate that neither Scp160 nor Bfr1 are essential for the recruitment of Dcp2 to ER membranes.

Bfr1 and Scp160 have been shown to be associated with polysomes (Lang et al., 2001), therefore, we asked whether P body components could be recruited to polysomes. We fractionated polysomes from wild-type, Δscp160, Δbfr1 and Δbfr1 Δscp160 cells on a sucrose gradient and subjected the polysome fractions to immunoblot analysis. As previously observed (Baum et al., 2004; Li et al., 2003), we did not find a significant change in the polysome profile or an increase in the 80S monosome fraction in any of the strains, which would be indicative of translation attenuation (Fig. 4A). In all cases, Dcp2–9Myc and Scd6–6HA were detected in the polysome fraction (Fig. 4A); thus, association of P body components with the polysomes is independent of Bfr1 and Scp160.

P bodies were not induced in the wild-type cells under the growth conditions that were used for the experiment; therefore, our data indicate that P body components are associated with polysomes during translation in normal logarithmically growing cells. We noticed that Dcp2–9Myc was detected as a double band by immunoblotting, and that, in the fractions of the Δscp160 Δbfr1 polysome profile, the upper band was strongly reduced. Thus, we tested whether the double band represented different phosphorylation states. Phosphorylation of Dcp2 has been previously reported, but no band shift in SDS gels had been observed (Yoon et al., 2010). Treatment of the gradient fractions with alkaline phosphatase caused the double band to collapse into the lower molecular weight band (Fig. 4B). These data suggest that phosphorylated and non-phosphorylated forms of Dcp2 are present on polysomes.

P bodies are present on ER membranes, and Scp160 and Bfr1 are enriched on ER-associated polysomes; yet, we used total yeast lysate for the polysome profiles. It is conceivable that P body components associate with the ER-bound polysomes in an Scp160–Bfr1-dependent manner. Moreover, the phosphorylated and dephosphorylated forms of Dcp2 might segregate between membrane-bound and soluble polysomes. To test these hypotheses, we recorded the polysome profiles of membrane and soluble fractions (supplementary material Fig. S1). Immunoblotting of the gradient fractions revealed that P body components could bind to soluble and membrane-associated polysomes (Fig. 4A). Interestingly, the ER-associated polysome pool was enriched for the phosphorylated form of Dcp2 (Fig. 4A). Phosphorylation of Dcp2 has been reported to be required for its accumulation in P bodies (Yoon et al., 2010), which is consistent with the notion that P bodies are associated with ER membranes (Kilchert et al., 2010). Notably, the amount of phosphorylated Dcp2 was reduced in the soluble polysome pool in Δscp160 Δbfr1 cells, indicating that Scp160 and Bfr1 together are required to maintain the Dcp2 phosphorylation state. Taken together, our data demonstrate that P body components are associated with polysomes and suggest that phosphorylated Dcp2 preferentially binds to membrane-associated polysomes.

The immunoelectron microscopy analysis revealed that, unlike in wild-type cells, in the Δscp160 and Δbfr1 strains, Dcp2–9Myc-positive structures were readily observed under normal non-stress growth conditions (Fig. 3A, 23°C). This observation led us to determine whether P bodies form in the mutant strains without any stress being applied at the normal growth temperature. Indeed, when we visualized Dcp2–GFP in the Δbfr1 and Δscp160 strains, and in the Δbfr1 Δscp160 double mutant, we found a large increase in the number of P bodies compared with the wild-type cells, which typically contained 0–2 P bodies per cell (Fig. 5A,B; supplementary material Fig. S2A). Importantly, this increase in Dcp2-positive structures was restricted to normal growth conditions. Shifting the mutant cells to 37°C, which causes a mild temperature stress – as indicated by the small increase in P body number in wild-type cells – reduced the number of the Dcp2-positive structures in the mutant strains to below even that of the wild type (Fig. 5A,B). We aimed to phenocopy the effect of the Δscp160 deletion by expressing an Scp160 C-terminal truncation that has been shown previously to compromise Scp160 function (Baum et al., 2004). The removal of the four most-C-terminal K homology (KH) domains induced the formation of multiple Dcp2-positive foci (Scp160 ΔC4, supplementary material Fig. S2B), which is consistent with this truncation having been shown to be hypersensitive to treatment with cycloheximide and the proposed role of Scp160 in translation (Baum et al., 2004). Deletion of either BFR1 or SCP160 causes a plethora of phenotypes; therefore, it is possible that the P body phenotype is caused by stress due to extended cultivation of the deletion strains. To rule out this possibility, we replaced the endogenous promoter of SCP160 and BFR1 with the inducible MET25 promoter, which is inhibited upon addition of methionine to the growth medium. Under these growth conditions, P body induction was still observed (supplementary material Fig. S2C). Taken together, these data suggest that Scp160 and Bfr1 negatively regulate P body formation under normal growth and stress-free conditions.

Given the increase of P bodies in Δbfr1 and Δscp160 cells and that, under certain conditions, P body formation is coupled to stress granule formation (Yoon et al., 2010), we next asked whether the number of stress granules is also increased under standard growth conditions in the mutant strains. Stress granule formation has been demonstrated to occur under glucose starvation and under heat stress (Buchan et al., 2008; Grousl et al., 2009). As a marker for stress granule formation, we used the polyU-binding protein Pub1p (Buchan et al., 2008). Indeed, we observed stress granule formation in wild-type cells under both conditions, whereas, in Δscp160 and Δbfr1 cells, stress granules were only visible after heat stress (Fig. 5C,D). Thus, Scp160 and Bfr1 might be required for stress granule formation only under certain stresses. Consistently, the Δbfr1 and Δscp160 strains did not induce stress granule formation under standard growth conditions (Fig. 5C,D); therefore, the phenotype of P body induction in the Δbfr1 and Δscp160 strains is not strictly coupled to stress granule formation.

Bfr1 and Scp160 form a complex at actively translating polysomes, and their localization at the ER is interdependent, nevertheless, they have been shown to have distinct functions (Lang et al., 2001). For example, only Scp160 binds mRNA, and only Bfr1 causes resistance to brefeldinA when overexpressed (Jackson and Képès, 1994; Lang and Fridovich-Keil, 2000; Weber et al., 1997). Moreover, Scp160, rather than Bfr1, bound to the mRNA associated P body components. We, therefore, wondered whether the interaction of Scp160 with Bfr1 is required to prevent P body formation under normal growth conditions, or whether Bfr1 and Scp160 can act independently. To distinguish between these possibilities, we overexpressed SCP160 and BFR1 by using the strong TEF1 promoter in the Δbfr1 and Δscp160 strains, respectively. Both genes were also overexpressed, individually, in the wild-type strain. In wild-type cells, the increased expression caused both proteins to also localize to the cytoplasm, indicating that the binding sites at the ER were saturable (Fig. 6A). Overexpression of BFR1 in Δscp160 cells, or SCP160 in Δbfr1 cells, reduced the number of Dcp2-positive foci to that found in wild-type cells (Fig. 6B,C). This reciprocal rescue was specific because overexpression of the unrelated ArfGAP GLO3, under the control of the TEF1 promoter, did not rescue the multiple-P-body phenotype in Δscp160 or Δbfr1 strains (supplementary material Fig. S2D); thus, merely increasing the mRNA levels did not rescue the P body phenotype. These data indicate that both Scp160 and Bfr1 are able to independently prevent P body formation under normal growth conditions.

So far we have shown that loss of Scp160 and Bfr1 induces Dcp2-positive foci. P bodies are highly dynamic structures in which P body components assemble in an aggregate-like manner, similar to prions. In fact, a number of P body proteins contain prion-like domains (Alberti et al., 2009; Reijns et al., 2008); thus, it would be conceivable that, in the absence of Scp160 or Bfr1, proper P bodies do not form. Given that the formation of P bodies requires the presence of RNA in these structures (Coller and Parker, 2005; Teixeira et al., 2005), we treated cells with cycloheximide, which ‘locks’ ribosomes onto mRNA and restricts the availability of the mRNA for the formation of P bodies. As expected, treatment of cells with cycloheximide abolished the Dcp2-positive foci in the mutant strains (Fig. 7A), indicating that Δscp160 and Δbfr1 cause the appearance of bona fide P bodies.

To expand our results, we determined the localization of three additional P body constituents (Scd6, Edc3 and Dhh1) (Fig. 7B and C). All of the markers formed multiple foci in Δbfr1 cells, only a few Scd6- or Edc3-positive structures were observed in Δscp160 cells and Dhh1 did not form foci that would correspond to P-bodies. To demonstrate that the foci formed by the different P body components associated with the same structures, we co-expressed Dcp2–GFP and Edc3–mCherry (supplementary material Fig. S2E). The localizations of the P body components showed extensive overlap, indicating that they are, indeed, present on the same structure. These data indicate that, in Δbfr1 cells, bona fide P bodies are formed in the absence of stress. By contrast, loss of Scp160 led to the formation of aggregates that lacked certain P body components, which we define as pseudo P bodies. Our results suggest that Scp160 might play a role in the assembly of P body proteins. Taken together, our data are consistent with both Bfr1 and Scp160 being negative regulators of P body formation in the absence of stress.

We hypothesized that because Bfr1 and Scp160 associate with actively translating polysomes, both proteins could prevent access of P body components to mRNA during translation. During stress conditions, under which translation is attenuated (Brengues et al., 2005; Kilchert et al., 2010), it is possible that polysomes disassemble and, therefore, that Scp160 and Bfr1 are lost from the mRNA, which means that they cannot protect mRNA and cannot be involved in the control of P body formation. Consequently, the loss of Scp160 or Bfr1 should not affect P body formation under such stress conditions. Upon withdrawal of glucose from wild-type, Δscp160 or Δbfr1 cultures, we observed P body induction to a similar extent in wild-type and Δbfr1 cells (Fig. 7D and E). A small increase in the number of Dcp2-positive structures was observed in the Δscp160 strain, consistent with a role for Scp160 in the correct assembly of P bodies. A similar result was obtained under heat stress (supplementary material Fig. S3). Our results are consistent with Scp160 and Bfr1 having a negative role on the formation of P bodies, but only in the absence of stress.

P body formation is an important process by which eukaryotic cells respond to a plethora of stresses, such as environmental, metabolic and cellular stress. Here, we aimed to understand why P bodies are localized in close proximity to the ER membrane. We find that P body components interact with the polysome-associated mRNA-binding protein Scp160. This interaction exposes the P body components to mRNA molecules as soon as they are released from the polysomes. Thus, the function of Scp160, and its interacting protein Bfr1, is, probably, to keep the P body components in a ‘waiting’ position but, nevertheless, inhibit the formation of P bodies under normal growth and stress-free conditions (Fig. 8). In this model, as soon as the cell encounters stress, the protective function of Scp160 and Bfr1 is lost, and P body components have unrestricted access to mRNAs.

We find that loss of either Bfr1 or Scp160 causes the formation of multiple Dcp2-positive structures. Although both structures contain mRNA, only the Dcp2-positive foci in Δbfr1 cells appear to be fully assembled P bodies (Fig. 8). In Δscp160 cells, the Dcp2-positive structures do not contain all of the P body components and, hence, are probably not functional in terms of mRNA decay. Thus, Scp160 appears to play a, perhaps, regulatory role in the correct assembly of P bodies. We assume that this role is more involved in the regulation of the process, because P bodies formed in Δscp160 strains under glucose starvation and also because stress granules appeared under heat stress.

The association of P body components with polysomes is, perhaps, not surprising. Different laboratories have previously proposed a model of dynamic equilibrium in which P body formation is directly coupled to translation attenuation (Coller and Parker, 2005; Franks and Lykke-Andersen, 2008). The most plausible way to achieve this coupling would be through a direct interaction. In addition, influencing the phosphorylation state of Dcp2 could be a part of the regulation of the dynamic equilibrium. Dcp2 is predominantly in the phosphorylated state on membrane-associated polysomes, whereas the phosphorylated and non-phosphorylated forms of Dcp2 are present in the soluble pool of polysomes. Phosphorylation of Dcp2 is required for the assembly of P bodies, which might influence the decay of some mRNAs (Yoon et al., 2010). Because P bodies are in close proximity to the ER, the phosphorylation state of Dcp2 might represent a ‘primed state’ for P body formation. Moreover, the P body components Dhh1 and Pat1 have been shown to be translational repressors, indicating that there might be a feedback mechanism between P bodies and ribosomes (Franks and Lykke-Andersen, 2008; Pilkington and Parker, 2008).

Scp160 and Bfr1 are, most likely, not the only proteins that can restrict the access of P body components to polysome-associated mRNA. Moreover, they are not essential for the association of P body components with polysomes, or the localization of P bodies close to the ER. Still, it is noteworthy that Scp160 is predominantly associated with membrane-bound polysomes (Weber et al., 1997). Thus, it cannot be excluded that Scp160 has some role in P body localization to the ER. It has been established that Scp160 only binds to a subset of mRNAs (Gelin-Licht et al., 2012; Hogan et al., 2008; Li et al., 2003), and a biological role for the mRNA–Scp160 interaction has been demonstrated for ASH1 mRNA localization and pheromone sensing (Gelin-Licht et al., 2012; Irie et al., 2002; Trautwein et al., 2004). Importantly, Scp160 has been shown to bind DHH1 mRNA and is required for efficient translation of this message (Li et al., 2003). Thus, Scp160 could be part of a regulatory loop that controls the translation of specific messages. In the case of Dhh1, however, we could not detect a difference in the protein levels in Δscp160 cells (supplementary material Fig. S4). This role could be regulated by phosphorylation because Scp160 is phosphorylated upon inhibition of the TOR complex by rapamycin, mimicking amino acid starvation (Soulard et al., 2010). In addition, Scp160 is part of a complex that contains Eap1, a negative regulator of translation (Cosentino et al., 2000; Mendelsohn et al., 2003). In fact, the Smy2, Eap1, Scp160, Asc1 (SESA) complex has been shown to act as a translational repressor of the POM34 mRNA in response to spindle pole body (SPB) duplication defects (Sezen et al., 2009). Pom34 is a component of the nuclear pore complex that plays a role in the insertion of the SPB into the nuclear envelope.

Scp160 appears to have multiple cellular functions – Scp160 controls ploidy (hence the name Saccharomyces protein controlling ploidy) (Wintersberger et al., 1995), it is required for the asymmetric localization of mRNAs (Gelin-Licht et al., 2012; Irie et al., 2002; Trautwein et al., 2004) and is part of a translational repressor complex (Sezen et al., 2009). Scp160 might also act as a signaling platform on ribosomes to control the translation of specific mRNAs (Baum et al., 2004), as it has been shown to act as an effector of a G-protein in the mating response pathway (Guo et al., 2003) and contributes to the regulation of telomere silencing (Marsellach et al., 2006). Furthermore, here, we show that Scp160 negatively regulates P body formation in the absence of stress and might be a novel P body component. How can all of these functions be performed by one protein? Scp160 could be part of different complexes that regulate Scp160 functions and recruit it to the correct site of function; such an adaptor function could be performed by Bfr1. Scp160 requires Bfr1 for efficient polysome association, and Bfr1 might recruit the SESA complex to the SPB (Sezen et al., 2009). However, Bfr1 must have additional functions, besides being an Scp160 interactor because overexpression of BFR1 in the Δscp160 strain was sufficient to alleviate the pseudo P body phenotype, and the concomitant loss of BFR1 and SCP160 showed a strong reduction of phosphorylated Dcp2 on soluble polysomes. Bfr1 might directly interact with P body components, although with a lower affinity than Scp160. Overexpression of BFR1 could, potentially, increase its interaction with P body components and maintain the P body components in the ‘waiting’ position.

In addition, the localization of ASH1 mRNA at the bud tip is not only dependent on Scp160 but also on Bfr1 and ribosomes (Irie et al., 2002; Trautwein et al., 2004). Thus, it is conceivable that Bfr1 is the adaptor for complexes that are involved in translation activation and repression. Clearly, Scp160 must have other adaptor proteins to fulfill its Bfr1-independent roles as a G protein effector or in the silencing of telomeres.

Scp160 is conserved in species up to human, and the closest homologues are vigilin (also known as HDLBP) and the fragile X syndrome protein FMRP (also known as FMR1) (Currie and Brown, 1999; Weber et al., 1997). Vigilin has been proposed to have a role in cytoplasmic mRNA metabolism in Drosophila and human cell lines (Goolsby and Shapiro, 2003) and to protect vitellogenin mRNA from degradation by the mRNA endonuclease PMR-1 in Xenopus (Cunningham et al., 2000). In addition, vigilin binds to the mRNA of the proto-oncogene c-fms and inhibits its translation, thereby relieving suppression of breast cancer cell proliferation (Cao et al., 2004; Woo et al., 2011). These functions might be regulated in a similar manner through binding to polysomes, because, akin to Scp160, vigilin binds to ribosomes through its C-terminal tail (Vollbrandt et al., 2004). Furthermore, the second mammalian homologue of Scp160 FMRP binds to a subset of mRNAs, and the I304N mutation, which was first described in an fragile X syndrome (FXS) patient, reduces the association of the protein with mRNAs and polysomes (Ascano et al., 2012; Feng et al., 1997; Siomi et al., 1994). Therefore, it is also conceivable that the function of Scp160 to control the translation of a subset of mRNAs is also conserved in vigilin and FMRP.

Standard genetic techniques were used throughout (Sherman, 1991). All modifications were carried out chromosomally, with the exceptions listed below. Chromosomal tagging and deletions were performed as described previously (Knop et al., 1999; Knop et al., 1999; Gueldener et al., 2002; Gueldener et al., 2002). For co-staining of the ER, the cells were transformed with the plasmid pSM1959 (with selectable marker LEU2) carrying Sec63–RFP (kindly provided by Susan Michaelis, Johns Hopkins University, Baltimore, MD). pRP1661 (with selectable marker URA3) encoding Pub1–mCherry and pRP1575 (with selectable marker TRP1) encoding Edc3–mCherry were provided by Roy Parker (University of Colorado, Boulder, CO). The plasmid pMS449 carrying the ΔC4 truncation of Scp160 was obtained from Matthias Seedorf (ZMBH Heidelberg, Germany). The plasmid pTEF-Glo3-FLAG (with selectable marker LEU2) was created through amplification of the GLO3 gene using the primers Glo3-6F and Glo3-11R (5′-GCAATTTAGGATCCAC-ATAGCGACAATGAG-3′ and 5′-GAGAAAATTCTCGAGATTTTCTTAAGTAG-3′, respectively) and the resulting fragment was cloned using BamHI and XhoI into pcDNA3.1 with a C-terminal FLAG tag.

The C-terminal FLAG tag was obtained through PCR of the plasmid pcDNA3.1+the Glo3-FLAG construct with the primers Glo3-6F and Glo3-10R (5′-GGCAAACTGCAGATGGCTG-3′). The resulting PCR fragment was digested with BamHI and PstI, and ligated into vector p425TEF. The final plasmids were confirmed by sequencing. The strains used in this study are listed in supplementary material Table S1.

The HBH purification was carried out as previously described by Tagwerker and colleagues (Tagwerker, et al., 2006) with modifications. Cells that expressed Dcp2–HBH or Scd6–HBH were grown to mid-to-late log phase at 23°C then shifted for 1 h to 37°C. The cells were fixed by adding 37% formaldehyde to a final concentration of 1% and incubating at room temperature for 2 min with gentle agitation. The formaldehyde was quenched for 5 min by the addition of glycine to a final concentration of 1.25 mM. The cells were harvested (4700 g for 3 min at 4°C), washed in 50 ml ice-cold double distilled H₂O, pooled, sedimented by centrifugation (5 min at 3000 g), flash frozen in liquid nitrogen and stored at −80°C. The cells were thawed and resuspended in 2 ml of buffer 1 (50 mM NaPi pH 8.0, 8M urea, 300 mM NaCl 20 mM imidazole, 0.5% Tween-20) per gram of yeast cell pellet. Per 2 ml screw capped tube containing 500 µl glass beads, 1 ml of the cell suspension was added. Lysis was performed in a FastPrep instrument four times for 45 seconds each at 4°C. The glass beads and cell debris were cleared with a slow spin (1300 g for 10 min at 4°C), and the supernatant was removed and pooled. The supernatants were spun at 20,000 g for 15 min at 4°C and the resulting pellets were resuspended in 500 µl buffer 1 with the addition of 1% Tween-20. 2.5 ml of Ni-NTA slurry that had been washed in buffer 1 was added to the pooled lysates and rotated overnight at room temperature. The tubes were spun at room temperature for 3 min at 2500 g and washed with 10 ml buffer 1. The tubes were spun as before and a series of 10 ml washes and spins were sequentially performed (buffer 2: 50 mM NaPi pH 6.4, 8 M urea, 300 mM NaCl, 0.5% Tween-20; buffer 3: 50 mM NaPi pH 8.0, 8 M urea, 300 mM NaCl, 0.5% Tween-20, 20 mM imidazole; buffer 4: 50 mM NaPi pH 8.0, 8 M urea, 300 mM NaCl, 0.5% Tween-20, 40 mM imidazole). After the final wash and spin, 4 ml of elution buffer (50 mM NaPi pH 6.4, 8 M urea, 300 mM NaCl, 0.5% Tween-20, 300 mM imidazole) was added and rotated with the beads for 45 min. The beads were spun down as before, the eluate was collected, and the elution step was repeated. 300 µl streptavidin agarose that had been washed in buffer 1 was added to the 8 ml of elute and rotated for 2 h at room temperature. The slurry was then spun as before and washed with 10 ml of buffer 6 (50 mM Tris-HCl pH 8.0, 8 M urea, 300 mM NaCl, 2% SDS). This step was then repeated with 10 ml buffer 7 (50 mM Tris-HCl pH 8.0, 8 M urea, 300 mM NaCl, 0.2% SDS). The final wash was performed in buffer 8 (50 mM Tris-HCl pH 8.0, 8 M urea, 300 mM NaCl) and the streptavidin agarose was resuspended in 1 ml buffer 8. The slurry was then washed three times with 1 ml 50 mM Tris-HCl pH 8.0 with 0.1% SDS, and 50 µl of solution was left in the tube after the final wash. To each tube, 0.25 µl of 1 mg/ml Endoproteinase LysC (ELC) was added, and the tubes were placed at 37°C with shaking. After 2 h, a further 0.25 µl of 1 mg/ml ELC was added and the tubes were placed at 37°C overnight with shaking. The tubes were briefly spun down and the supernatant was collected. The beads were extracted twice with of 50 mM Tris-HCl pH 8.0 with 0.1% SDS (in a volume of 200 µl followed by 100 µl). The three samples were pooled and dried in a Speed-Vac concentrator. The peptides were resuspended in 400 µl BS (85% acetonitrile, 10 mM ammonium acetate with formic acid pH 3.5) and loaded twice on a HILIC TopTip PolyLC spin column (PolyLC) that had been equilibrated three times with 50 µl ES (5% acetonitrile, 10 mM ammonium acetate with formic acid pH 3.5) and three times with 50 µl BS. The column was washed three times with 50 µl BS and the peptides were eluted four times with 25 µl ES. All elutions were pooled and dried in a Speed-Vac concentrator. The peptides were then resuspended in 50 µl of 50 mM NH₄HCO₃ and subjected to trypsin digest with 0.5 µl 1 mg/ml trypsin (Promega) for 6 h at 37°C. Mass spectrometric analysis was performed using LC-MS/MS (Orbitrap).

Cells were grown in yeast extract peptone dextrose (YPD) medium to early-log phase and then either shifted to 37°C for 1 h or subjected to various stresses, as indicated. The cells were spun at 3000 g for 3 min and resuspended in Hartwell's complete (HC) medium and immobilized on concanavalin-A-coated slides. For strains harboring the pTEF-Glo3-FLAG plasmid, the cells were grown in HC medium that lacked leucine. In the methionine experiments, cells were grown at 30°C in YPD containing 100 µg/ml methionine for 72 h and maintained at an OD₆₀₀ of between 0.2 and 1.0 through continuous dilution.

Fluorescence was monitored with an Axiocam mounted on an Axioplan 2 fluorescence microscope (Carl Zeiss, Oberkochen, Germany) by using Axiovision software. Image processing was performed using Image J and Adobe Photoshop CS5 (San Jose, CA). For counting, a minimum of 200 cells from at least three independent experiments was counted per condition. In the quantification graphs, the size of the box is determined by the 25th and 75th percentiles, the whiskers represent the 5th and 95th percentiles, the lines and the triangles mark the median and the mean, respectively.

Cells expressing Dcp2–9Myc were grown to early-log phase at 23°C and then shifted to 37°C for 1 h. The cells were fixed and treated for immunoelectron microscopy as described previously (Prescianotto-Baschong and Riezman, 2002). Secondary antibodies against goat anti-rabbit IgG that had been coupled to 10-nm gold particles (BBInternational, Cardiff, United Kingdom) were used to detect the binding of polyclonal rabbit antibodies that recognise Myc (Abcam, Cambridge, MA).

Immunoprecipitations were performed similarly as described previously (Sezen et al., 2009). Twenty OD₆₀₀ units of logarithmically growing cells that had been grown in YPD and subjected to the various stresses, as indicated, were disrupted with glass beads in 300 µl of immunoprecipitation buffer [50 mM triethanolamine pH 6.0, 150 mM KCl, 5 mM EDTA, 5 mM EGTA, protease inhibitor tablet (Roche), 1 mM PMSF and 1 mM benzamidine]. The lysates were incubated with 1% Triton X-100 for 10 min at 4°C, and the extracts were cleared by centrifugation (3000 g for 10 min at 4°C). An aliquot was removed that served as the input control, and the remainder of the extract was incubated with a monoclonal antibody against FLAG peptide (M2, Invitrogen) and magnetic bead slurry (Dynal) for 2.5 h at 4°C. The beads were washed three times with immunoprecipitation buffer containing 0.1% Triton X-100. The beads were resuspended in sample buffer and analyzed by immunoblotting.

For the RNase A treated samples, after the extracts had been cleared by centrifugation, they were treated with 40 U of RNase inhibitor (Promega) or 1 µg/µl RNase A and incubated for 1 h at 4°C. The extracts were then spun for 3 min at 3000 g, and the supernatant was used for the immunoprecipitation as described above.

Antibodies were used to detect the following epitope tags and proteins: Myc (9E10, 1∶1000, Sigma-Aldrich), hemagglutinin (HA11, 1∶1000, Covance, Princeton, NJ), Scp160 (1∶1000), Bfr1 (1∶1000), Rpl16a (1∶1500, a gift from Michael N. Hall, Biozentrum, Basel, Switzerland) and Sec61 (1∶10,000, a gift from Martin </emph>Spiess, Biozentrum, Basel, Switzerland).

Flotation of ER membranes was performed as described previously (Schmid et al., 2006). The equivalent of 50 OD₆₀₀ units was converted into sphereoplasts at 37°C and lysed by Dounce homogenization in 3 ml lysis buffer [20 mM HEPES in KOH, pH 7.6, 100 mM sorbitol, 100 mM KAc, 5 mM Mg(Ac)₂, 1 mM EDTA, 1 mM dithiothreitol, 100 µg/ml cycloheximide, protease inhibitors]. After the removal of cellular debris (5 min, 300 g), the membranes were pelleted by centrifugation (10 min, 13,000 g), resuspended in 2 ml lysis buffer containing 50% sucrose and layered on top of 2 ml 65% sucrose in lysis buffer. Two additional 5 ml and 2 ml cushions (40% and 0% sucrose) were layered on top. The step gradient was spun in a TST41.14 rotor for 16 h at 28,000 g. After centrifugation, 1 ml fractions were collected from each of the cushions and the interfaces, and were then precipitated with trichloroacetic acid (TCA). The samples were analyzed by SDS-PAGE and immunoblotting.

Preparations of polysomes were performed as described previously (de la Cruz et al., 1997) on 4–47% sucrose gradients that had been prepared using a Gradient Master (Nycomed Pharma, Westbury, NY). Gradient analysis was performed using a gradient fractionator (Labconco, Kansas City, MO) and the Äcta FPLC system (GE Healthcare) and continuously monitored at A₂₅₄. Fractions were precipitated with 10% TCA, separated using SDS-PAGE and analyzed by western blotting.

We thank Cornelia Kilchert (Biozentrum, University Basel), Roy Parker, Matthias Seedorf, Michael N. Hall, Martin Spiess, Paul Jenö, Suzette Moes, Ian G Macara for reagents. Paul Jenö and Suzette Moes (Biozentrum, University of Basel) are acknowledged for the mass spectrometric analysis of the HBH purifications. Ian G. Macara is acknowledged for critically reading the manuscript.