The precise functions of most of the ∼200 assembly factors and 79 ribosomal proteins required to construct yeast ribosomes in vivo remain largely unexplored. To better understand the roles of these proteins and the mechanisms driving ribosome biogenesis, we examined in detail one step in 60S ribosomal subunit assembly—processing of 27SA3 pre‐rRNA. Six of seven assembly factors required for this step (A3 factors) are mutually interdependent for association with preribosomes. These A3 factors are required to recruit Rrp17, one of three exonucleases required for this processing step. In the absence of A3 factors, four ribosomal proteins adjacent to each other, rpL17, rpL26, rpL35, and rpL37, fail to assemble, and preribosomes are turned over by Rat1. We conclude that formation of a neighbourhood in preribosomes containing the A3 factors establishes and maintains stability of functional preribosomes containing 27S pre‐rRNAs. In the absence of these assembly factors, at least one exonuclease can switch from processing to turnover of pre‐rRNA.
Formation of mature ribosomal subunits involves interplay between folding, modification, and processing of pre‐rRNA, and binding of ribosomal proteins (r‐proteins) to pre‐rRNA (reviewed in Sykes and Williamson, 2009). Biogenesis of ribosomal subunits in Saccharomyces cerevisiae also requires the activity of ∼200 transiently associating assembly factors (reviewed in Henras et al, 2008; Kressler et al, 2009). These factors bind preribosomes at specific steps of the assembly pathway, perform their functions, and eventually dissociate before the formation of mature ribosomal subunits. The precise mechanisms by which these assembly factors and r‐proteins exert their roles during ribosome assembly are only now beginning to be understood.
Initially, studies of ribosome assembly in yeast focused on identification of intermediates in pre‐rRNA processing, discovery of assembly factors, and determining which factors were required for which steps in pre‐rRNA processing or nuclear export of pre‐rRNPs. Subsequently, physical and genetic interactions between factors were identified, assembly intermediates were purified and their constituents determined, and assembly subcomplexes were discovered (reviewed in Henras et al, 2008; Kressler et al, 2009). It appears that most assembly factors have been found and initially assigned to one or more steps in subunit biogenesis. Thus, it is now possible to more comprehensively examine how all factors known to participate in one step work together to drive each assembly step and the biogenesis pathway. Such focused approaches have provided more detailed insights into mechanisms of late steps in maturation of pre‐40S and pre‐60S particles (reviewed in Panse and Johnson, 2010).
To better understand the mechanism of biogenesis of 60S subunits in yeast, we have focused on the pre‐rRNA processing step involving the exonucleolytic removal of ITS1 sequences of 27SA3 pre‐rRNA to form 27SB1S pre‐rRNA (‘A3 processing step’) (Figure 1A and C). Proper processing of 27SA3 pre‐rRNA is important because it generates the 5′‐end of the major form of mature 5.8S rRNA, 5.8SS rRNA. Furthermore, this processing event may initiate a conformational switch necessary to form functional ribosomes. Secondary structure models predict that ITS1 sequences in 27SA3 pre‐rRNA, removed by the A3 processing step, basepair with sequences in what will become the 5′‐end of 5.8SS rRNA (Yeh et al, 1990; van Nues et al, 1994). In mature ribosomes, the same sequences of 5.8SS rRNA basepair with 25S rRNA and provide a binding site for the r‐protein rpL17 (Taylor et al, 2009; Ben‐Shem et al, 2010; Supplementary Figure S1). Exonucleolytic trimming of 27SA3 pre‐rRNA to 27SB1S pre‐rRNA therefore provides an excellent framework to understand how the interplay between RNA folding, RNA processing, and protein binding enables maturation of preribosomes.
Approximately 80 different assembly factors have been found in one or more 66S preribosomes that are precursors to mature 60S ribosomal subunits. Effects on 60S subunit biogenesis have been examined for 70 of these 80 factors. A defect in 27SA3 pre‐rRNA processing is evidenced by the accumulation of the 27SA3 pre‐rRNA precursor and reduction of the immediate downstream product, 27SB1S pre‐rRNA (Dunbar et al, 2000; Pestov et al, 2001; Adams et al, 2002; Gadal et al, 2002; Oeffinger et al, 2002; Oeffinger and Tollervey, 2003; Horsey et al, 2004; Miles et al, 2005; Figure 1B; Supplementary Table S3). Thus far, mutation or depletion of only seven of these 70 assembly factors tested unambiguously results in a pre‐rRNA processing phenotype diagnostic of a role in 27SA3 pre‐rRNA processing. These proteins, which we refer to as the ‘A3 factors’, include potential scaffolding proteins (Nop7, Erb1, and Ytm1), RNA‐binding proteins (Rlp7 and Nop15), and proteins with no predicted function (Nsa3/Cic1, Rrp1).
The 5′–3′ exonucleases Rat1, Xrn1, and Rrp17 together with Rai1 (Rat1‐interacting protein) also are required for processing 27SA3 pre‐rRNA, and normally halt precisely at the B1S site (Henry et al, 1994; Xue et al, 2000; Oeffinger et al, 2009). When 60S subunit assembly is aborted upon depletion of A3 factors, 27S pre‐rRNAs undergo turnover (Dunbar et al, 2000; Pestov et al, 2001; Adams et al, 2002; Gadal et al, 2002; Oeffinger et al, 2002; Oeffinger and Tollervey, 2003; Horsey et al, 2004; Miles et al, 2005). Interestingly, Rat1 has been shown to function in the turnover of poly(A)+ 27S pre‐rRNAs in strains defective for the exosome (Fang et al, 2005). This result led us to ask whether Rat1 might have a role in turnover of defective 27S pre‐rRNAs generated in A3 factor mutants.
In order to understand the events underlying the A3 processing step, and to understand the principles involved in eukaryotic ribosome biogenesis, we have addressed the following issues: (1) the timing of association of A3 factors with preribosomes, (2) the interdependence among A3 factors for association with preribosomes, (3) the role of A3 factors, if any, in recruitment of exonucleases to preribosomes, and in stopping exonucleases precisely at the B1S site, (4) whether A3 factors enable association of other assembly factors and r‐proteins with preribosomes, which might be required for subsequent remodelling events within preribosomes, and (5) the fate of preribosomes in A3 factor mutants, specifically whether they are turned over by the Rat1 exonuclease.
A3 factors associate with preribosomes in a concerted manner, well before processing of 27SA3 pre‐rRNA
To begin to understand how 27SA3 pre‐rRNA processing occurs, we asked when and how all assembly factors required for this processing step are recruited into preribosomes. Nop7 and Erb1 first associate with 90S preribosomes containing 35S pre‐rRNA, whereas Rrp1 and Nop15 associate subsequently with 66S preribosomes containing 27SA2 pre‐rRNA (Oeffinger and Tollervey, 2003; Horsey et al, 2004; Miles et al, 2005; Figure 1C). In order to determine the timing of entry of the remaining A3 factors—Rlp7, Nsa3/Cic1, and Ytm1, we assayed with which pre‐rRNA species each protein copurifies. The 90S factor Enp1 was used as a positive control for coimmunoprecipitation (co‐IP) of 35S pre‐rRNA. Co‐IP of 35S pre‐rRNA revealed that, like the other A3 factors, Rlp7, Nsa3/Cic1, and Ytm1 assemble early into preribosomes, well before their requirement for processing of 27SA3 pre‐rRNA (Figure 1C and D).
Since A3 factors are required for the same pre‐rRNA processing step and some of them physically interact with each other (Miles et al, 2005; Tang et al, 2008), we examined whether they are interdependent for their association with preribosomes. We purified preribosomes from strains in which each A3 factor was depleted and asked what happened to the other A3 factors. Because all seven assembly factors are essential, we engineered conditional GAL promoter strains to regulate expression of each gene (Longtine et al, 1998). We used TAP‐tagged assembly factor Rpf2 to purify preribosomes, since Rpf2 is present in all precursors to 60S subunits and does not have a known role in processing of 27SA3 pre‐rRNA (Zhang et al, 2007). Preribosomes were purified from strains grown in galactose and after shifting from galactose to glucose‐containing medium for 16 h to deplete each assembly factor.
As shown before, Nop7, Erb1, and Ytm1 are interdependent for assembly into preribosomes (Figure 2A; Tang et al, 2008). Silver staining and mass spectrometry also revealed that Rlp7, Nop15, and Nsa3/Cic1 were significantly decreased in preribosomes lacking Nop7, Erb1, or Ytm1 (Figure 2A; unpublished observations). Likewise, upon depletion of Rlp7, Nop15, or Nsa3/Cic1, the five other A3 factors (except Rrp1) were greatly diminished in preribosomes (Figure 2A and B; Supplementary Figure S2A). The SDS–PAGE profiles of preribosomes upon depletion of each of these six interdependent proteins were remarkably similar (Figure 2A), and appeared to be unique to factors required for 27SA3 pre‐rRNA processing. Very different changes in the composition of Rpf2‐containing preribosomes were seen in strains mutant for any of 20 other assembly factors that function in different steps of pre‐rRNA processing, for example upon depletion of Rea1 (Figure 2A; unpublished observations). Therefore, the observed gel profiles of preribosomal proteins in A3 factor mutants are specific and not simply a secondary effect of prolonged growth in glucose‐containing medium, to deplete the A3 factors.
Western blotting of preribosomes purified from the GAL–RLP7 strain is shown here as an example of interdependent association of A3 factors (Figure 2B). In the absence of Rlp7, levels of the A3 factors Nop7, Ytm1, Nsa3/Cic1, and Nop15 are diminished in preribosomes (Erb1 was not tested). Although not in preribosomes under these conditions, the three proteins Nop7, Erb1, and Ytm1 are present as a heterotrimeric subcomplex (Miles et al, 2005) (Figure 2C). In contrast, factors required for other steps in pre‐rRNA processing and assembly were unchanged when Rlp7 was depleted (Figures 2A and B and 3A and B; Supplementary Figure S2B). When Rrp1 was depleted from cells, few changes in preribosome composition could be detected by inspection of stained proteins resolved by SDS–PAGE (Figure 2D).
Thus, six A3 factors—Nop7, Erb1, Ytm1, Rlp7, Nop15, and Nsa3/Cic1 are interdependent for their association with preribosomes, whereas Rrp1 appears to associate independently. Owing to the interdependence of A3 factors (Figure 2), all experiments described henceforth were performed only in the GAL–RLP7 strain.
To more thoroughly assay global changes in the composition of preribosomes upon depletion of A3 proteins, we took a proteomic approach, using iTRAQ (Ross, 2004). We compared preribosomes purified from the GAL–RLP7 RPF2–TAP strain grown in galactose or shifted to glucose, to identify proteins that either increased or decreased in the absence of Rlp7 (Figure 3A and B; Supplementary Figure S2B; Supplementary Table S1). We identified 177 proteins from 1859 spectra (⩾95% confidence), including 51 assembly factors (Figure 3A), 63 r‐proteins (Figure 3B), and 62 other proteins (Supplementary Figure S2C).
Levels of ∼15 assembly factors associated with preribosomes were increased upon depletion of Rlp7 (Figure 3A; see also Figure 2A). These assembly factors function in early steps of ribosome biogenesis, before processing of 27SA3 pre‐rRNA (Dragon et al, 2002; reviewed in Henras et al, 2008; Kressler et al, 2009). Their accumulation in preribosomes when 27SA3 pre‐rRNA processing is blocked is in agreement with the observed enrichment of early pre‐rRNAs in these mutants. Consistent with the SDS–PAGE and immunoblot analysis (Figure 2A and B), levels of Nop7, Nop15, and Nsa3/Cic1 in preribosomes were reduced in the absence of Rlp7. Erb1 and Ytm1 were not identified by iTRAQ, as is often the case in our experience. Assembly factors Nog2, Nsa2, and Alb1 also were reduced, consistent with their joining preribosomes after completion of 27SA3 pre‐rRNA processing (Saveanu et al, 2001, 2003; Lebreton et al, 2006). Cgr1, Spb1, Ybl028c, Nop16, and Dbp10 proteins exhibited the greatest reduction in preribosomes when A3 factors were absent. However, it is not known when these assembly factors are present in preribosomes. While Dbp10 and Spb1 function after the A3 processing step (Kressler et al, 1999; Burger et al, 2000; Moy et al, 2002; Fleischer et al, 2006), mutants for Cgr1, Nop16, and Ybl028c have not been thoroughly assayed. Therefore, it remains possible that one or more of these assembly factors might be as yet unidentified A3 factors. Levels of most other assembly factors did not change, consistent with SDS–PAGE profiles and western blotting (Figures 2A and B and 3A; Supplementary Figure S2B).
The 5′–3′exonuclease Rrp17, but not Rat1 or Xrn1, requires A3 factors for association with preribosomes
Since the exonucleases Rat1, Xrn1, and Rrp17 are required for processing 27SA3 pre‐rRNA, we hypothesized that A3 factors might recruit these enzymes to preribosomes. Therefore, in the absence of A3 factors, these exonucleases would no longer be present in preribosomes, resulting in accumulation of unprocessed 27SA3 pre‐rRNA. iTRAQ and western blotting revealed that this was the case for Rrp17 (Figures 3A and 4A). In contrast, western blotting revealed that the remaining two exonucleases, Rat1 and Xrn1, still were able to associate with preribosomes in the absence of Rlp7 (Figure 4A). Therefore, we conclude that the absence of Rrp17 is responsible, at least in part, for the accumulation of 27SA3 pre‐rRNA observed in A3 factor mutants.
Rat1 is not directed to preribosomes by its cofactor Rai1 and enters preribosomes before creation of its substrate, the 5′‐end of 27SA3 pre‐rRNA
Since Rat1 appears to be a major exonuclease involved in 27SA3 pre‐rRNA processing (Henry et al, 1994), but is not recruited to preribosomes by A3 factors, we wanted to explore other mechanisms by which Rat1 might be recruited to preribosomes. Rai1 physically interacts with Rat1 (Xue et al, 2000), and stimulates its exonuclease activity (Xiang et al, 2009). Therefore, we investigated whether Rai1 directs Rat1 to preribosomes, by assaying levels of Rat1 in preribosomes purified from the rai1Δ strain. Rat1 was present in these preribosomes, indicating that Rai1 is not required for assembly of Rat1 into preribosomes (Figure 4B).
To test whether Rat1 is recruited to preribosomes by its RNA substrate, the 5′‐end of 27SA3 pre‐rRNA, we assayed preribosomes when cleavage at the A3 site is blocked. To do so, we depleted Pop3 (Dichtl and Tollervey, 1997), a protein component of RNase MRP that cleaves 27SA2 pre‐rRNA at the A3 site (Lygerou et al, 1996). Rat1–HA was still present in preribosomes in the absence of Pop3, suggesting that the 5′‐end of 27SA3 pre‐rRNA does not have to be generated for Rat1 to be recruited to preribosomes (Figure 4C).
The results above suggest that Rat1 may be present in preribosomes before the A3 end is generated. Therefore, we assayed with which pre‐rRNAs Rat1 is associated. Primer extension revealed that Rat1 co‐IPed with large amounts of 27SA2 pre‐rRNA (Figure 4D), indicating that it joins preribosomes before generation of 27SA3 pre‐rRNA. Therefore, Rat1 might be recruited to preribosomes by other early associating assembly factors.
rpL17, rpL26, rpL35, and rpL37 cannot stably associate with preribosomes lacking A3 factors
Of the 63 r‐proteins identified in preribosomes by iTRAQ, 18 were from the 40S ribosomal subunit (rpS proteins). The levels of most rpS proteins were significantly increased in Rpf2‐containing preribosomes in the absence of Rlp7. This is consistent with the presence of 40S subunit r‐proteins in 90S preribosomes, and the accumulation of early ribosomal intermediates including 90S preribosomes when 27SA3 pre‐rRNA processing is blocked. In contrast, r‐proteins of the 60S subunit were affected to varying extents. Some were unaffected, whereas others were modestly reduced (Figure 3B). Notably, four r‐proteins were significantly diminished in preribosomes when Rlp7 was depleted—rpL17, rpL26, rpL35, and rpL37 (Figure 3B). Immunoblotting confirmed this result from iTRAQ experiments (Figure 5A). These four r‐proteins were also diminished when two other A3 factors, Nop7 and Nop15, were depleted (Supplementary Figure S4).
Based on the crystal structure of the yeast 60S ribosomal subunit (Ben‐Shem et al, 2010), rpL17, rpL26, rpL35, and rpL37, lie close to each other, adjacent to 5.8SS rRNA, whose 5′‐end is generated by 27SA3 pre‐rRNA processing (Figure 5B and C; Supplementary Figure S1). rpL17 is particularly interesting because it contacts helix 2, formed between the 5′‐end of 5.8SS rRNA and 25S rRNA. We imagined that binding of rpL17 to helix 2 might form an RNP structure that stops Rat1 precisely at the B1S site. Therefore, we wanted to understand when and how rpL17 associates with preribosomes.
rpL17 becomes more stably associated with preribosomes after processing of 27SA3 pre‐rRNA
The absence of rpL17 from preribosomes in A3 factor mutants might indicate that it associates with pre‐rRNPs only after processing of 27SA3 pre‐rRNA is completed. Therefore, in mutants where 27SA3 pre‐rRNA processing is blocked, rpL17 might not yet have assembled into nascent ribosomes. Alternatively, rpL17 might associate with preribosomes before the 27SA3 pre‐rRNA processing step, but its binding may be strengthened by rRNP remodelling. In A3 factor mutants, where such rearrangements might not occur, weakly bound rpL17 might readily dissociate from preribosomes. To begin to distinguish between these possibilities, we determined the timing of association of rpL17 with nascent ribosomes, by assaying with which pre‐rRNA processing intermediates it co‐IPs (Figure 6A). rpL17 co‐IPed 27SA2 and 27SB pre‐rRNAs, and a small amount of 35S pre‐rRNA (Figure 6A). No 27SA3 pre‐rRNA was detected copurifying with rpL17, most likely due to the very low abundance of this pre‐rRNA species in cells under wild‐type conditions.
To further investigate the timing and degree of association of rpL17 with preribosomes, we measured its copurification with ‘early’ (Npa2) or ‘middle’ (Nsa1, Nog2) assembly intermediates. Consistent with experiments above, we found that rpL17 is present in early preribosomes purified using the TAP‐tagged early assembly factor Npa2 (Figure 6B). Npa2 is thought to leave the assembly pathway during or immediately after processing of 27SA2 pre‐rRNA (Rosado et al, 2007). However, reproducibly more rpL17 copurified with later assembly factors such as Nsa1 or Nog2 (Figure 6B and C) (Saveanu et al, 2003; Kressler et al, 2008). No such difference in copurification could be observed for the other r‐protein tested, rpL8 (Figure 6B and C). Taken together, the above results suggest that rpL17 normally associates with preribosomes before processing of 27SA3 pre‐rRNA. Thus, there are two possible explanations for the reduced association of rpL17 in A3 factor mutants. On the one hand, rpL17 might completely fail to associate with preribosomes in the absence of A3 factors. Alternatively, rpL17 might be weakly bound to preribosomes early in the assembly pathway, and may dissociate from pre‐rRNPs if there is a block in 27SA3 pre‐rRNA processing.
The consistently elevated amounts of rpL17 observed in Nsa1 and Nog2‐containing preribosomes, compared with Npa2‐containing pre‐rRNPs, suggests that rpL17 is in fact more tightly associated with later preribosomes, than with early pre‐rRNPs. We therefore tested salt sensitivity of association of rpL17 with preribosomes. Association of rpL17 with Npa2–TAP, but not Nog2–TAP, was salt sensitive (Figure 6C), indicating that association of rpL17 with later preribosomes is stronger than with early preribosomes. No such difference in association was observed for the other r‐protein tested, rpL8 (Figure 6C). The difference in recovery of rpL17 in preribosomes at 50 mM NaCl compared with higher concentrations of salt is not merely an artefact of contamination with mature ribosomes, since levels of rpL8 that copurify with preribosomes remain constant at different salt concentrations (Figure 6C). Taken together, these results suggest that rpL17 assembles into early preribosomes before the 27SA3 pre‐rRNA processing step, but its association is strengthened after subsequent remodelling events, coincident with production of 27SB1S pre‐rRNA.
We hypothesized that exonucleolytic processing of 27SA3 pre‐rRNA might be the event that triggers more stable association of rpL17 with preribosomes. To test this idea, we used the GAL–RRP17 MET–RAT1 xrn1Δ strain, in which all three exonucleases required for 27SA3 pre‐rRNA processing can be conditionally repressed (Oeffinger et al, 2009). Because A3 factors remained in preribosomes when Rat1, Xrn1, and Rrp17 were depleted, using this strain allowed us to prevent removal of ITS1 sequences in 27SA3 pre‐rRNA, in a manner independent of the A3 factors (Figure 6D). However, equal amounts of rpL17 could still associate with preribosomes (Figure 6D), indicating that 27SA3 pre‐rRNA processing is not required for strengthening association of rpL17. It is possible that rpL17 is stably bound to helix 2, which might have already formed before 27SA3 pre‐rRNA processing occurs. The above result also suggests that the stable association of rpL17 with preribosomes depends on the presence of A3 factors or some other remodelling event induced by their presence.
rpL17 is required for processing within ITS2
The presence of rpL17 near the 5′‐end of helix 2 in mature ribosomes suggests that it could serve as a roadblock to Rat1, causing it to stop precisely at the B1S site during 5′–3′ processing of ITS1. This idea is further supported by the results above that rpL17 is associated with preribosomes before processing of 27SA3 pre‐rRNA. To test whether rpL17 functions to halt Rat1, we determined the effect of depleting rpL17 on pre‐rRNA processing. In the absence of rpL17, we observed accumulation of pre‐rRNA species whose 5′‐ends are ∼10 nts downstream of the B1S site, very close to the end of helix 2 (Figure 7, left). This 5′‐truncated pre‐rRNA has not been observed in the absence of any other assembly factor or r‐protein tested, including rpL35 (Zhang et al, 2007; Pöll et al, 2009; Babiano and de la Cruz, 2010; Supplementary Figure S3). This shorter pre‐rRNA may form due to the inability of Rat1 to stop at the B1S site in the absence of rpL17. To test whether Rat1 is required for the formation of this shorter pre‐rRNA species, we constructed a temperature‐sensitive rat1‐1 strain where rpL17 was depleted using the GAL promoter. When rpL17 was depleted and Rat1‐1 was inactivated by shifting to 37 °C, the 5′‐truncated product observed in the absence of rpL17 alone no longer accumulated (Figure 7, right), revealing a direct role for Rat1 in the formation of these pre‐rRNAs. Therefore, it appears that in a small fraction of preribosomes lacking rpL17, Rat1 starts processing at the A3 site, but cannot stop at the B1S site. Instead, it is stopped ∼10 nts downstream of this site.
The significance of this 5′‐truncated pre‐rRNA species, however, is unclear, as it appears that the major fraction of 27SA3 pre‐rRNA is properly processed by the exonucleases to form 27SB1S pre‐rRNA. We observed an increase in 27SB1S pre‐rRNA (Figure 7, left), indicating that rpL17 is required for the next step in pre‐rRNA processing—conversion of the 27SB1S pre‐rRNA to 25.5S and 7S pre‐rRNAs. Consistent with these data, it has been shown that there is a reduction in the amount of 7S pre‐rRNA in the absence of rpL17 (Pöll et al, 2009).
Rat1 initiates turnover of misassembled preribosomes when A3 factors are absent
Pulse‐chase and steady‐state assays of pre‐rRNA processing in A3 factor mutants suggest that 27S pre‐rRNA is turned over when 60S subunit assembly is aborted (Dunbar et al, 2000; Pestov et al, 2001; Adams et al, 2002; Gadal et al, 2002; Oeffinger et al, 2002; Oeffinger and Tollervey, 2003; Horsey et al, 2004; Miles et al, 2005). Because Rat1 is present in preribosomes when A3 factors are depleted (Figure 4A), we hypothesized that Rat1 might degrade the 27SA3 pre‐rRNA in abortive preribosomes, instead of properly processing it. Thus, turnover intermediates degraded from their 5′‐ends might be detectable. Primer extension assays revealed the presence of such turnover intermediates. We detected low levels of pre‐rRNA species with 5′‐ends 67–69 nts downstream of the B1S site, when Rat1 was present in preribosomes lacking A3 factors (Figure 8A, lanes 1–12).
To confirm that Rat1 is required for formation of these shorter RNAs when A3 factors are absent, we constructed a strain in which both Rlp7 and Rat1 were depleted. In this strain, no 5′‐truncated pre‐rRNAs accumulated (Figure 8B, lanes 1 and 2). Similar results were obtained upon depletion of Rlp7 in a temperature‐sensitive rat1‐1 strain or rat1D235A catalytic mutant strain (Figure 8B, lanes 1–4). Taken together, these results indicate a direct role for Rat1 exonuclease activity in turnover of 27S pre‐rRNA in A3 factor mutants.
The work described here has enabled us to bring into sharper focus the assembly factors known to be involved in the processing of 27SA3 pre‐rRNA. At the same time, we have also been able to make connections between these assembly factors and r‐proteins in a manner that is consistent with the structure of S. cerevisiae 60S ribosomes.
We have shown that all known A3 factors associate with preribosomes well before their requirement for processing of 27SA3 pre‐rRNA. Six of seven A3 factors assemble into preribosomes in a concerted manner. A3 factors enable stable association of one of the 5′–3′ exonucleases, Rrp17, but not Rat1 or Xrn1, with preribosomes. Four ribosomal proteins—rpL17, rpL26, rpL35, and rpL37 specifically cannot associate with preribosomes when A3 factors are depleted. These four r‐proteins bind adjacent to each other on 5.8S rRNA in mature 60S ribosomes in S. cerevisiae (Ben‐Shem et al, 2010). This result indicates that the presence of A3 factors, which are required for proper formation of the 5′‐end of 5.8SS rRNA, stabilizes this neighbourhood of r‐proteins within assembling ribosomes. We also show here that in the absence of A3 factors and rpL17, rpL26, rpL35, and rpL37, Rat1 cannot stop at the B1S site and proceeds beyond this site to turn over 27S pre‐rRNA (Figure 8C).
Recruiting functions of A3 factors during ribosome assembly
The seven assembly factors required for processing of 27SA3 pre‐rRNA enable stable association of three classes of proteins with preribosomes—other A3 assembly factors, an exonuclease, and r‐proteins. First, A3 factors ensure productive assembly of other A3 factors into preribosomes. Such a pattern of interdependence could reflect both direct and indirect interactions—directly by protein–protein contacts, and indirectly via a network of RNA–protein and RNA–RNA contacts. Indeed, recent proteome‐wide protein complementation analysis has revealed that these proteins, except Rrp1, are in the vicinity of each other (Tarassov et al, 2008). Consistent with these data, Nop7, Erb1, and Ytm1 form a stable heterotrimer that can be isolated from preribosomes (Miles et al, 2005; Figure 2C). It has been proposed that this Nop7 subcomplex might serve as a scaffold upon which further assembly can occur (Lapik et al, 2004). CRAC analysis (Granneman et al, 2009) revealed that Nop7 and Erb1 bind RNA sequences that lie close to each other in 25S rRNA near the ITS2 junction, whereas Nop15 and Cic1 bind sequences within ITS2 (Granneman et al, 2011). As Ytm1 has been shown to interact directly with Erb1 (Miles et al, 2005), it follows that its binding site on preribosomes is very close to those of Erb1 and Nop7. Since these five A3 factors are potentially present in such close proximity to each other in preribosomes, one can well imagine how their association with preribosomes would be interdependent.
Second, A3 factors are required for the association of Rrp17, one of the three 5′–3′ exonucleases that processes 27SA3 pre‐rRNA to 27SB1S pre‐rRNA (Oeffinger et al, 2009). Thus, the absence of Rrp17 could account for the observed accumulation of some unprocessed 27SA3 pre‐rRNA in A3 factor mutants. Two‐hybrid assays revealed no physical interactions between any of the A3 factors and Rrp17 (unpublished observations), suggesting that A3 factors might affect association of Rrp17 with preribosomes indirectly through their effect on the structure of nascent ribosomes. Interestingly, it appears that the association of the remaining exonucleases, Rat1 and Xrn1, is not affected in A3 mutants by such changes in preribosomal architecture. They might be anchored to preribosomes or properly positioned within pre‐rRNPs by other assembly factors and/or r‐proteins that remain stably associated with preribosomes in the absence of A3 factors.
A third important function of A3 factors is to ensure stable association of r‐proteins rpL17, rpL26, rpL35, and rpL37 with preribosomes. Interestingly, the binding sites in mature ribosomes of rpL17, rpL26, rpL35, and rpL37 are near each other in domains I and III of 5.8S/25S rRNAs (Ben‐Shem et al, 2010). This appears to be close to the binding sites of Erb1 in domain I, Nop7 in domain III, and potentially Nop15 and Nsa3/Cic1 in ITS2 within preribosomes (Granneman et al, 2011). We have not detected physical interactions between A3 factors and these r‐proteins (unpublished observations), suggesting that binding of these r‐proteins to preribosomes also depends on remodelling or stabilization of preribosome structure mediated by A3 factors. It is conceivable that the binding of Nop7 and Erb1 to domains I and III of 5.8S/25S rRNAs, respectively, might help bring these two rRNA domains in close proximity, thereby providing binding sites for the four r‐proteins and stabilizing their association with preribosomes.
Rat1 might function as a processing or degradative enzyme depending on the presence or absence of roadblocks
It has been observed that in mutants defective in ribosome biogenesis, pre‐rRNA processing intermediates do not accumulate to high levels, suggesting that when assembly fails, pre‐rRNAs are turned over. Mistakes made in the course of normal ribosome assembly in wild‐type cells may also trigger turnover. In fact, in wild‐type cells, polyadenylation targets pre‐rRNAs for 3′–5′ degradation by the exosome (LaCava et al, 2005). In strains defective for the exosome, depletion of Rat1 results in increased accumulation of poly(A)+ pre‐rRNAs (Fang et al, 2005), suggesting that the 5′–3′ exonuclease activity of Rat1 plays a role in turnover of aberrant pre‐rRNAs during ribosome assembly. Our work provides evidence for the role of Rat1 in 5′–3′ degradation of specific aberrant pre‐rRNA intermediates when ribosome assembly is aborted in A3 factor mutants. This is evident by the accumulation of shorter Rat1‐dependent RNAs in the absence of A3 factors and rpL17 (Figures 7 and 8A). Consistent with this idea, there is greater accumulation of 27SA3 pre‐rRNA in the GAL–RLP7 rat1‐1 double mutant compared with the GAL–RLP7 single mutant (Figure 8B). However, this Rat1‐mediated turnover might not be a very efficient process, since in the absence of A3 factors, significant amounts of 27SA3 intermediate still can be detected relative to the short degradation products. It is also likely that turnover by Rat1 is not the only pathway for degradation of aberrant pre‐rRNA intermediates.
An important question in ribosome biogenesis is why cells employ both endonucleases and exonucleases for pre‐rRNA processing. The exonucleases Rat1 and Xrn1 are implicated in generating precise 5′‐ends of two mature rRNAs—5.8SS and 25S rRNAs (Geerlings et al, 2000). While at first glance the use of an exonuclease to create precise 5′‐ends of rRNA might appear counterintuitive, it is in fact an ingenious display of cellular economy. Exonucleases, such as Rat1, could process pre‐rRNA to create appropriate 5′‐ends when stopped by roadblocks such as RNA secondary structure or RNA–protein contacts present in preribosomes. However, under conditions of aberrant ribosome assembly as signalled by the absence of the roadblock, the same exonuclease might not stop at a specific site to create a precise 5′‐end. Instead, its processing function would now be converted to a turnover function, without necessitating recruitment of other factors explicitly for the purpose of turnover. Such a dual role has recently been proposed for the mammalian homologue of Rat1, Xrn2, during 60S ribosome biogenesis (Wang and Pestov, 2011).
rpL17 binds to the 5′‐end of 5.8S rRNA (Ben‐Shem et al, 2010), immediately downstream of the B1S site where exonucleolytic processing precisely halts in wild‐type cells. rpL17 is present in preribosomes before processing of 27SA3 pre‐rRNA is completed. We therefore predicted that rpL17 might serve as a roadblock to the exonucleases Rat1, Xrn1, and Rrp17 during processing of 27SA3 pre‐rRNA, causing them to stop at the B1S site. Our results indicate that in part rpL17 is a roadblock to Rat1 (Figure 7), as seen by the small fraction of Rat1‐dependent 5′‐truncated pre‐rRNAs observed in the absence of rpL17. However, rpL17 most likely does not act alone to stop Rat1. Instead, it is likely that an RNP structure created by the stable association of A3 factors and all four r‐proteins with preribosomes might serve as the roadblock to Rat1.
A3 factors may be necessary to help fold ITS2 RNA, and bring together 5.8S and 25S rRNA sequences to stabilize 27S pre‐rRNAs during ribosome biogenesis
Because all seven A3 factors are present in preribosomes after 27SA3 pre‐rRNA processing is completed (Figure 1C), they might also function in later steps of ribosome biogenesis, for example processing of 27SB pre‐rRNA into 25.5S plus 7S pre‐rRNAs. This is thought to be the last step that occurs before nucleolar release of pre‐60S particles (Kressler et al, 2009). Consistent with this idea, preribosomes are found to remain in the nucleolus in ytm1‐1, cic1‐2, and rlp7‐1 mutants (Gadal et al, 2002; Fatica et al, 2003; Miles et al, 2005). Furthermore, the AAA+‐ATPase Rea1 releases the Nop7–Erb1–Ytm1 subcomplex from preribosomes just prior to the exit of pre‐60S particles from the nucleolus (Baßler et al, 2010). Therefore, A3 factors might also play a structural role to enable proper processing of 27SB pre‐rRNA, before they are released from 66S preribosomes.
Indeed, three observations suggest a role for A3 factors in facilitating proper folding and processing of ITS2, and thus contributing to 27SB1S pre‐rRNA processing. (1) Nop15 and Nsa3/Cic1 bind ITS2 sequences, whereas Nop7 and Erb1 bind 25S rRNA sequences close to the ITS2‐25S junction (Granneman et al, 2011). (2) Cleavage at the C2 site within ITS2 occurs prematurely in the cic1‐2 mutant (Fatica et al, 2003). (3) Mutations in Pes1 and Bop1, the mammalian homologues of Nop7 and Erb1, respectively, slow processing of ITS2 (Strezoska et al, 2000; Lapik et al, 2001).
Three of the four r‐proteins most affected in the A3 mutants are not required for 27SA3 pre‐rRNA processing, but are required for 27SB pre‐rRNA processing. In the absence of rpL17, rpL35 or rpL37, 27SB pre‐rRNA, but not 27SA3 pre‐rRNA accumulates (Babiano and de la Cruz, 2010; M Gamalinda, personal communication). Consistent with these observations, A3 factors are present in preribosomes even in the absence of rpL17, rpL35, or rpL37 (M Gamalinda, personal communication). The block in processing of 27SB pre‐rRNA indicates that these three r‐proteins are required to establish or maintain structures required for processing of 27SB pre‐rRNA, but not 27SA3 pre‐rRNA. Recently, we also found that rpL17 is required for the stable association of several assembly factors necessary for proper processing of 27SB to 7S + 25.5S pre‐rRNAs (M Gamalinda, personal communication).
Binding of A3 factors to domains I and III of rRNA sequences in preribosomes, coupled with the assembly of the r‐protein neighbourhood adjacent to 5.8S rRNA, might therefore be required for maintaining the preribosomal architecture necessary for proper processing of 27SA3 pre‐rRNA and stabilization of 27SB1S pre‐rRNA, before cleavage and removal of ITS2 (Figure 8C). We propose that the A3 factors and r‐proteins facilitate folding of 5.8SS and 25S rRNA sequences in pre‐rRNAs, including formation of helices 2, 4, and 10 between them (Supplementary Figure S1), and influence proper folding of ITS2 within 27S pre‐rRNAs. Proper maintenance of these helices is likely required not just for precise processing of 27SA3 pre‐rRNA, but also for the correct processing of later intermediates such as 27SB1S, 7S, and 25.5S pre‐rRNAs. Once A3 factors dissociate from preribosomes, it is imperative that this basepairing between 5.8S and 25S rRNAs be maintained. Stable association of rpL17, rpL26, rpL35, and rpL37 may play a role in maintaining basepairing between these two rRNAs after release of A3 factors, and in mature functioning ribosomes (Supplementary Figure S1).
Materials and methods
3xHA‐, GFP‐, 13xMYC‐, TAP‐tagged genes, GAL1 promoter fusions, and gene disruptions were generated as described in Longtine et al (1998). Yeast strains used in this study are listed in Supplementary Table S1.
Single‐step purifications of preribosomes from whole‐cell extracts using magnetic Dynabeads were performed as described in Oeffinger et al (2007), with the following modifications. Cells were grown in either galactose, to express GAL promoter fusion genes, or shifted to glucose for 16 h to repress these genes, and harvested at mid‐log phase (3 × 107 cells/ml). Cultures were centrifuged using a GS‐3 rotor at a speed of 5000 r.p.m., for 5 min. Cell pellets were resuspended in RNP buffer (20 mM HEPES (pH 7.4), 110 mM KOAc, 0.5% Triton, 0.1% Tween 20, 40 mM NaCl), and subjected to glass bead lysis. After binding of lysates to IgG‐coated Dynabeads at 4 °C for 30 min, beads were washed three times with RNP buffer to eliminate non‐specifically bound proteins. Bead‐bound preribosomes were eluted by cleaving the TEV protease site within the TAP‐tag, using 10 U of TEV Protease (Invitrogen). Proteins were recovered from the eluate by precipitation in 10% TCA, and were subsequently suspended in SDS sample buffer and separated by SDS–PAGE on 4–20% polyacrylamide NOVEX gels (Invitrogen). For purifications done under high salt conditions, the following modifications were made. Cell lysis was performed using RNP buffer containing 40 mM NaCl. After binding of lysates to Dynabeads, three washes were done with RNP buffer containing the indicated concentration of NaCl. These were followed by three washes with RNP buffer containing 40 mM NaCl to get rid of excess salt that might affect the efficiency of TEV Protease.
Silver staining was done according to standard procedure. Proteins present in whole‐cell extracts or purified preribosomes were assayed by western blot analysis (Ausubel et al, 1994), with the following modification. After electroblotting, the nitrocellulose membrane was cut into smaller sections based on known mobility of different proteins to enable probing of multiple proteins from one blot. TAP‐tagged proteins were detected using alkaline phosphatase conjugated to IgG (Pierce). 3HA‐tagged proteins were identified with mouse monoclonal antibody 12CA5, GFP‐tagged proteins with rabbit polyclonal anti‐GFP antibody, and myc‐tagged proteins with anti‐mouse 9e10 antibody. Otherwise, antibodies specific for ribosomal proteins or assembly factors were used.
Proteins copurifying with TAP‐tagged Rpf2 were identified by mass spectrometry as described in Horsey et al (2004). For semi‐quantitative iTRAQ analysis, preribosomes were purified as described above. Following TCA precipitation, pellets were resuspended in 20 μl of 20 mM HEPES pH 7.4. iTRAQ labelling and quantification were done as described in Jiang et al (2007). The hash that represents these data is iCHfI8YFKDk4nxNV79z9Z+AcPF8B6tPwYtsNRV8FK+Dq7qQPZpgLpS0bbrEfsgaxlmjUBK6pnxlLoDZjbjGQkK9Kfv8AAAAAACipjQ==, located at the network ProteomeCommons.org Tranche, under the title ‘Assembly of Saccharomyces cerevisiae 60S Ribosomal Subunits: Role of Factors Required for 27S Pre‐rRNA Processing’.
RNA from whole‐cell lysates was extracted as described in Horsey et al (2004). RNA enriched from purified preribosomes was extracted as follows. After binding of cell‐free lysate to magnetic beads, the bead‐bound preribosomes were treated with 5 μl of Proteinase K (Roche) for 30 min at 37 °C to degrade proteins. RNA was extracted from the eluate using phenol chloroform isoamyl alcohol. Primer extension was carried out using 32P‐ radiolabelled oligonucleotide sequences complementary to 27S pre‐rRNA (Miles et al, 2005).
PyMOL images of ribosome structure were generated using PDB File 3O58 (Ben‐Shem et al, 2010).
Supplementary data are available at The EMBO Journal Online (http://www.embojournal.org).
Conflict of Interest
The authors declare that they have no conflict of interest.
Source data for Figures 1A, 1D, 6A, 6B, 7, 8B
We thank the following people for generously providing strains and plasmids, and antibodies against the indicated proteins: Steven Buratowski (RAT1 and rat1D235A plasmids), Marlene Oeffinger (GAL–RRP17 MET–RAT1 xrn1Δ strain), Brooke McCartney (GFP), Adam Linstedt (Myc), David Goldfarb (Nip7), Michael McAlear (Ebp2), Micheline Fromont‐Racine (Nog2 and Nsa2), François Lacroute (rpL1), Jon Warner (rpL3), Arlen Johnson (rpL8, rat1‐1 strain), Juan Pedro Ballesta (rpL12), Sabine Rospert (rpL17), K Siegers (rpL25), Maurice Swanson (rpL39), and Elisabeth Tosta (Cic1). We thank Susan Dowd for assistance with mass spectrometry. We are grateful to Dimitri Pestov and members of our laboratory for critical reading of the manuscript.
Author contributions: The iTRAQ experiment and data analysis were done by JS, JM, and PA. Primer extension assays of A3 factor mutants and two‐hybrid assays were done by JD. All other experiments were performed by AS, and designed and analysed by AS and JLW. The manuscript was written by AS and JLW, with input from JD, JS, PA, and JRM. This work was supported by NIH Grants GM28301 to JLW, and GM77628 to JRM.
- Copyright © 2011 European Molecular Biology Organization