Previous Article | Next Article ![]()
Eukaryotic Cell, December 2004, p. 1619-1626, Vol. 3, No. 6
1535-9778/04/$08.00+0 DOI: 10.1128/EC.3.6.1619-1626.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Department of Molecular Biophysics and Biochemistry,3 Department of Genetics,1 Department of Therapeutic Radiology, Yale University School of Medicine, New Haven, Connecticut2
Received 11 August 2004/ Accepted 15 September 2004
|
|
|---|
|
|
|---|
![]() View larger version (20K): [in a new window] |
FIG. 1. Schematic diagram of pre-rRNA processing in S. cerevisiae. The 35S pre-rRNA is transcribed as a single transcript that is subsequently cleaved at the A0, A1, and A2 sites by the SSU processome. Cleavage at A2 or A3 separates precursors to the small-subunit (18S) and large-subunit (5.8S and 25S) rRNAs, whereas cleavage at A0 or A1 matures the 5' end of the 18S rRNA. The broken line represents multiple processing steps. The locations of oligonucleotides a, b, c, e, and y, used to probe for precursor and mature rRNAs in subsequent experiments, are shown.
|
The U3 snoRNA and its associated proteins are required for the processing of the small ribosomal subunit at cleavage sites A0, A1, and A2 (Fig. 1). Cleavages at A0 and A1 in the 5' external transcribed spacer mature the 5' end of the pre-rRNA. Cleavage at A2 or A3 in internal transcribed spacer 1 separates the small-ribosomal-subunit precursor rRNA from the large-ribosomal-subunit precursors. Defects in cleavage at the A0, A1, and A2 sites lead to a reduction in the levels of the 18S rRNA. This reduction causes accumulation of the 35S and 23S pre-rRNAs and a reduction in the levels of the 27SA2 and 20S pre-rRNAs (39).
A large RNP required for the processing of the small-ribosomal-subunit rRNA, called the small-subunit (SSU) processome, has recently been purified (4). This preribosomal complex contains the U3 snoRNA and at least 28 proteins. We defined the SSU processome components as having the following properties. (i) They are nucleolar. (ii) They are able to coimmunoprecipitate with the U3 snoRNA and Mpp10 (a protein specific to the SSU processome). (iii) They are required for 18S rRNA biogenesis. Subsequent large-scale tandem affinity purification studies have also purified several 80 to 90S preribosomal complexes that contain many SSU processome components (9, 33). Additional proteomic studies have also revealed similar proteins required for ribosome biogenesis and subcomplexes, including SSU processome proteins (17, 29). Collectively, these studies have purified an 80S or 90S preribosome which serves as a precursor to the 43S preribosome and is required for cleavages at A0, A1, and A2. Here, we present results that expand upon these studies and validate the role of seven nonribosomal proteins in pre-rRNA processing.
During the original purification, analysis was limited to proteins whose peptide sequences were present more than once in the mass spectrometric analysis (4). Due to this stringent criterion, we hypothesized that there may be some additional SSU processome components that had been eliminated in the original purification. In addition, the initial purification included the presence of five ribosomal proteins (Rps4, Rps6, Rps7, Rps14, and Rps28), but it was not known at that time whether they were components of the SSU processome or contaminants. Therefore, we aimed to determine whether these five ribosomal proteins are SSU processome components by testing (i) their association via coimmunoprecipitation with known SSU processome components, i.e., Mpp10 and the U3 snoRNA, and (ii) their association with precursors to the 18S rRNA. Since there is no in vitro ribosome assembly system in yeast, the order of assembly of ribosomal proteins in eukaryotes is currently unknown. We have found a subset of ribosomal proteins to be associated with the SSU processome, suggesting that these proteins are able to associate with early precursors to the 18S rRNA. Therefore, the association of specific ribosomal proteins with the SSU processome suggests that it has roles in both pre-rRNA processing and ribosome assembly.
We have also tested whether seven other nonribosomal proteins (Utp18, Noc4, Utp20, Utp21, Utp22, Emg1, and Krr1), present only once in the purifications or subsequently found by others to coimmunoprecipitate with SSU processome components, were additional components of the SSU processome (7, 8, 32). A subset of these proteins has been partially characterized as being involved in ribosome biogenesis (3, 7, 9, 17, 25, 29, 33). For example, Utp18 was first identified (9), localized to the nucleolus (12), and shown to coimmunoprecipitate with the 5' external transcribed spacer and the U3 snoRNA (17). Here we show that Utp18 is required for pre-18S rRNA processing by Northern blot analysis of pre-rRNAs; we also show that Utp18 coimmunoprecipitates with the SSU processome protein, Mpp10, by Western blot analysis. Similarly, Noc4 was previously identified (9, 29) and localized (9, 26), and the pre-rRNA processing phenotype of the temperature-sensitive Noc4 mutant was analyzed (26). However, Noc4 was found not to coimmunoprecipitate with the U3 snoRNA (9). Here, we determined the resulting defects in pre-rRNA processing of cells depleted of Noc4 and report that, contrary to previously published results, Noc4 does indeed coimmunoprecipitate with both the U3 snoRNA and the SSU processome protein, Mpp10. In this study, we determined that these proteins are components of the SSU processome, which therefore places them in a specific preribosomal complex. Much to our surprise, we found that the SSU processome associates with the 23S pre-rRNA, a precursor that has been cleaved at A3 but not at A0, A1, or A2. Furthermore, we are able to unify, support, and expand upon previous studies. In addition to further characterizing the nonribosomal proteins, we have validated that five ribosomal proteins, previously not known to be components of pre-rRNA complexes, are components of the SSU processome.
Collectively, we have validated that these 12 proteins are indeed bona fide SSU processome components and are essential for 18S rRNA biogenesis. Together, these results suggest that the SSU processome, in addition to its role in pre-rRNA processing, also has a role as a ribosome assembly intermediate.
|
|
|---|
63 his3-
200 leu2-
1). Yeast strains were grown in rich medium, either YEPD (1% yeast extract, 2% peptone, 2% glucose) or YPG/R (1% yeast extract, 2% peptone, 2% galactose, 2% raffinose) as specified below. Expression of proteins from a conditional promoter. Strains which expressed N termini that were tagged with a triple hemagglutinin epitope tag (3xHA) from a galactose-inducible and glucose-repressible promoter were created as described previously (22) with plasmid pFA6a-kanMX6-PGAL1-3xHA and with primers with 50 nucleotides of complementarity to the gene of interest (the Utp1, Utp7, Utp18, Noc4, Utp20, Emg1, Bfr2, Enp1, Rps9B, or Rps14A gene).
C-terminal 3xHA tagging. Yeasts expressing proteins with 3xHA tags were constructed as described previously (15) with plasmid pYM1 (kanMX6 selectable marker) and 50 nucleotides with complementarity to the gene of interest (the Utp1 to Utp10, Utp12 to Utp17, Utp21, Utp22, Krr1, Enp1, Enp2, Rpf2, Imp4, Rps4A, Rps6B, Rps7B, Rps27A, Rps28A, Rpl33A, Nop7, or Sof1 genes).
Analysis of pre-rRNA processing by Northern blotting. Strains expressing an N-terminal 3xHA tag were grown in YPG/R and then washed and resuspended in YEPD. RNA was extracted from 10 ml of cells grown to an optical density at 600 nm of 0.4 to 0.5 in YPG/R and of cells grown in YEPD (24 h). RNA extraction and Northern blotting were carried out as previously described (20). Equal amounts of RNA (5 µg) were loaded in each lane.
Immunoprecipitations. Immunoprecipitations were carried out with N- and C-terminally 3xHA-tagged strains. Protein-protein coimmunoprecipitations were carried out with 200 µl each of anti-HA (12CA5 hybridoma culture supernatant) with glass bead protein extracts and blotted with anti-Mpp10 antibodies (20). Immunoprecipitations for protein, RNA, and pre-rRNA were carried out with extracts from strains that had been N-terminally tagged with GAL-3xHA and C-terminally tagged with 3xHA as previously described.
Protein-snoRNA coimmunoprecipitations were carried out with 200 µl of anti-HA (12CA5) on tagged protein extracts made by glass bead disruption, and RNA was extracted and analyzed as previously described (20). Protein-pre-rRNA coimmunoprecipitations were carried out as previously described (40), except that Rps4A and Rps6B pre-rRNA coimmunoprecipitates were probed with oligonucleotides e and b.
Immunofluorescence. Yeast strains expressing 3xHA-tagged proteins were used in indirect immunofluorescence assays to determine subcellular localization as described previously (1, 5). Mouse anti-HA (12CA5 hybridoma culture supernatant; dilution, 1:1,000) and rabbit anti-Mpp10 polyclonal antibodies (dilution, 1:2,000) (6) were detected with tetramethylrhodamine B isothiocyanate-conjugated goat anti-mouse immunoglobulin G (dilution, 1:100) and fluorescein isothiocyanate-conjugated donkey anti-rabbit immunoglobulin G (dilution, 1:300) secondary antibodies (Jackson ImmunoResearch). The localization of tagged Utp4, Utp18, Noc4, Utp20, Bfr2, and Emg1 was carried out with N-terminally 3xHA-tagged strains, whereas the localization of Utp21, Utp22, and Enp2 was carried out with C-terminally 3xHA-tagged strains.
|
|
|---|
During the original SSU processome purification, five copurifying ribosomal proteins, Rps4, Rps6, Rps7, Rps14, and Rps28, were identified (4). In order to assess whether these and other ribosomal proteins might be components of the SSU processome, 3xHA-tagged ribosomal proteins Rps4, Rps6, Rps7, Rps9, Rps14, Rps27, Rps28, and Rpl33 were tested for their ability to coimmunoprecipitate with Mpp10 and the U3 snoRNA, two components of the SSU processome (Fig. 2). Four of the five SSU processome-copurifying ribosomal proteins (Rps4, Rps6, Rps7, and Rps14) were able to coimmunoprecipitate with Mpp10 and the U3 snoRNA, though to different degrees. Rps9 was also able to coimmunoprecipitate with Mpp10 and the U3 snoRNA. Relative to the other ribosomal proteins, Rps7 was able to coimmunoprecipitate only with small amounts of Mpp10 and the U3 snoRNA. In contrast, ribosomal proteins Rps27, Rps28, and Rpl33 did not coimmunoprecipitate with Mpp10 or the U3 snoRNA (Fig. 2). As expected, no coimmunoprecipitation was observed with the untagged parental strain, YPH499, while Utp9 (a known SSU processome component) did coimmunoprecipitate with both Mpp10 and the U3 snoRNA (4). We verified that each ribosomal protein (both positive and negative) was enriched by immunoprecipitation by stripping the blot and reprobing it for the HA-tagged protein (data not shown).
![]() View larger version (60K): [in a new window] |
FIG. 2. A subset of ribosomal proteins is found in the SSU processome. 3xHA-tagged ribosomal proteins Rps4, Rps6, Rps7, Rps9, Rps14, Rps27, Rps28, and Rpl33 were immunoprecipitated with beads conjugated with HA antibodies. 3xHA-tagged Utp9 (a bona fide SSU processome component) and YPH499 (the untagged parental strain) were used as positive and negative controls, respectively. Ribosomal proteins were tested with Western blotting for their abilities to coimmunoprecipitate Mpp10. Ribosomal proteins were also tested with Northern blotting for their abilities to coimmunoprecipitate the U3 snoRNA. Results for totals (lanes T), representing 5% (Western blot) and 10% (Northern blot) of the proteins extracted, and immunoprecipitates (lanes IP) are shown.
|
|
View this table: [in a new window] |
TABLE 1. Additional nonribosomal components of the SSU processome
|
|
View this table: [in a new window] |
TABLE 2. Nucleolar proteins that are not components of the SSU processome
|
![]() View larger version (28K): [in a new window] |
FIG. 3. Subcellular localization of new SSU processome proteins and two other related proteins. All proteins were 3xHA tagged and colocalized using anti-HA(red) or nucleolar protein Mpp10 (green) with anti-Mpp10 by indirect immunofluorescence. DAPI (4',6'-diamidino-2-phenylindole) (blue) was used to stain the nuclear DNA. Merge, merge of tagged protein, Mpp10, and DAPI. (A) Nucleolar localization of Utp18, Noc4, Utp20, Utp21, Utp22, and Emg1. The parental strain YPH499 (untagged) and Utp4 (a bona fide SSU processome component) were used as negative and positive controls, respectively. (B) Nucleolar localization of Bfr2 and Enp2.
|
![]() View larger version (64K): [in a new window] |
FIG. 4. Coimmunoprecipitation experiments define new proteins as SSU processome components. Immunoprecipitations of 3xHA-tagged Utp18, Noc4, Utp20, Utp21, Utp22, Emg1, Krr1, and Enp1 were carried out using anti-HA antibodies. 3xHA-tagged proteins were immunoprecipitated and tested for their ability to coimmunoprecipitate with Mpp10 as determined by Western blot analysis. U3 snoRNA that coimmunoprecipitated with 3xHA-tagged proteins were analyzed by Northern blotting. Results for totals (lanes T), representing 5% (Mpp10) or 10% (U3 snoRNA) of the total protein extracted, and for immunoprecipitates (lanes IP) are shown. The parental strain, YPH499 (untagged), and 3xHA-tagged Rpf2 (a protein involved in large-subunit biogenesis) (40) were used as negative controls. 3xHA-tagged Utp7 (a bona fide SSU processome component) was used as a positive control.
|
![]() View larger version (116K): [in a new window] |
FIG. 5. Depletion of new SSU processome proteins leads to defects in pre-18S rRNA processing. Strains expressing Utp1, Utp7, Utp18, Noc4, Utp20, Emg1, Bfr2, and Enp1 from galactose-inducible and glucose-repressible promoters were grown in early log phase (time point 0) in galactose- and raffinose-supplemented media (undepleted) and then shifted into glucose (depleted) for 24 h. YPH499, the untagged parental strain, was used as a control. RNA from undepleted or depleted yeast strains was analyzed for the presence of pre-rRNAs by Northern blotting. Equal amounts of RNA were separated on formaldehyde-1.2% agarose gels, transferred to membranes, and hybridized with the specific oligonucleotide probes shown in Fig. 1. (A) Northern blot using oligonucleotide c, which hybridizes to 35S, 32S, 27SA2, 23S and 21S pre-rRNAs; (B) Northern blot using oligonucleotides b and e, which hybridize to the 35S, 32S, 27SA, 27SB, 23S, 20S pre-rRNAs; (C) Northern blot using oligonucleotides a and y, which hybridize to the 18S and 25S rRNAs, respectively.
|
![]() View larger version (66K): [in a new window] |
FIG. 6. SSU processome and small-ribosomal-subunit proteins associate with precursor rRNAs. 3xHA-tagged Imp4, Utp1, Utp4, Utp9, Utp16, Utp17, Utp20, Utp21, Rps4, Rps6, and Rpf2 were immunoprecipitated with anti-HA antibodies bound to Sepharose beads and analyzed by Northern blotting for their ability to coimmunoprecipitate with pre-rRNAs (lanes HA). An untagged strain, YPH499, and 3xHA-tagged Rpf2, a protein involved in large-subunit biogenesis, were used as controls. Imp4, a known SSU processome component, was used as a positive control. All experiments were also done in parallel without antibody, i.e., with beads alone (lanes BA), and pre-rRNA precursors that immunoprecipitated were determined by comparison to a standard (lane STD). Northern blots were probed with oligonucleotides b, c, and e, which hybridize to the 35S, 32S, 27SA2, 27SB, 23S, 21S, and 20S pre-rRNAs (as shown in Fig. 1). Rps4 and Rps6 Northern blots were probed with oligonucleotides b and e, which hybridize to the 35S, 32S, 23S, 21S, and 20S pre-rRNAs (see Fig. 1).
|
|
|
|---|
The SSU processome is thus a preribosomal complex of at least 40 proteins and the U3 snoRNA. As might be expected based on the large number and sizes of the proteins, this complex sediments at 80S on a sucrose gradient (4). Surprisingly, pre-rRNA coimmunoprecipitation experiments demonstrate that SSU processome proteins efficiently associate with the 23S pre-rRNA, a precursor which is often labeled as an aberrant pre-rRNA (2, 39). However, this precursor is normally present in the strain in our study (10). Therefore, the SSU processome remains associated with rRNA precursors of the mature 18S rRNA that have been separated from rRNA precursors to the large ribosomal subunit by cleavage at A3.
Although we tested a large number of ribosomal and nonribosomal proteins to find additional components of the SSU processome, we may not have exhausted all possible candidates. We tested a subset of small-ribosomal-subunit proteins based upon their identification in our original purification of the SSU processome and their identification in other purifications (4, 8, 9). Although we found a distinct set of ribosomal proteins to be components of the SSU processome, there may be additional ribosomal protein components that were not tested. In addition, we may not have identified some nonribosomal protein components because we relied on the existing copurifications and cannot rule out the possibility that the conditions used did not disrupt the integrity of this complex. Currently, the SSU processome is thought to be composed of the U3 snoRNA, 35 nonribosomal proteins (Utp1 to Utp18, Utp20 to Utp22, Noc4, Nop1, Nop56, Nop5/58, Snu13, Mpp10, Imp3, Imp4, Dhr1, Rrp9, Rrp5, Emg1, and Krr1), and 5 ribosomal proteins (Rps4, Rps6, Rps7, Rps9, and Rps14) (4).
The order of assembly of ribosomal proteins with rRNA was first described for Escherichia coli during the early 1970s (27, 35). The first steps in analysis of the ribosomal pattern of assembly came in 1966, when Staehelin and Meselson observed that 30 to 40% of the small-ribosomal-subunit proteins partially disassembled during density gradient centrifugation in 5 M cesium chloride (35). This discovery enabled the establishment of a system for reconstituting ribosomes, which facilitated the elucidation of a detailed pathway for in vitro ribosome assembly, termed the "30S assembly map" (11, 27, 28, 36). Ribosomal proteins were grouped according to their abilities to bind to rRNA and to each other. Primary binders (i.e., S4, S7, S8, S15, S17, and S20) are ribosomal proteins that bind to rRNA directly, whereas secondary and tertiary binders are ribosomal proteins that require the presence of one or more ribosomal proteins (28). Many of the bacterial primary binding proteins (for example, S7, S8, S15, and S20) do not have yeast homologues, making it difficult to extrapolate the bacterial data to S. cerevisiae (23, 30). Only Rps9 and Rps14 yeast ribosomal proteins have bacterial homologues, i.e., the primary and tertiary binding proteins S4 and S11, respectively (23, 30). Since S. cerevisiae rRNAs and bacterial rRNAs are different, and since not all ribosomal proteins are conserved in both organisms, S. cerevisiae may have a set of primary binding proteins that is distinct from that in bacteria.
We propose that the ribosomal proteins associated with the SSU processome may be analogous to the primary or secondary binding proteins described for bacteria, since cleavages by the SSU processome represent early pre-rRNA maturation steps for the small-ribosomal-subunit rRNA. We found that the yeast ribosomal proteins Rps4, Rps6, Rps7, Rps9, and Rps14 were bona fide components of the SSU processome and may therefore represent a distinct set of yeast ribosomal proteins involved in the early stages of ribosome assembly. Because there is no in vitro ribosomal assembly system for eukaryotic ribosomes, we can only hypothesize which ribosomal proteins bind first on the basis of their association with pre-rRNAs. For example, Rps4 and Rps6 were both able to coimmunoprecipitate the 23S pre-rRNA, suggesting that they may be involved in ribosome assembly prior to cleavage at sites A0, A1, and A2 (Fig. 6). These results are consistent with those of Kruiswijk et al., who hypothesized that a specific set of ribosomal proteins (Rps23, Rps18, Rps2, Rps30, Rps5, Rps11, Rps19, Rps4, Rps21, Rps9, Rps22, and Rps3 [in new nomenclature]) were involved in the early stages of ribosomal assembly (18). In agreement with this hypothesis, we found that Rps4 and Rps9 may be required for the early steps of ribosome assembly. However, we were unable to confirm the current ribosomal protein counterpart for 10 ribosomal proteins (S19, S12, S22, S20, S21, S11, S6, S17, S5, and S29) that were found by Kruiswijk et al. to be associated with an early step of assembly of the small ribosomal subunit (18, 23, 30). In addition, two ribosomal proteins that we have found to be SSU processome components, Rps7 and Rps14, were not analyzed by Kruiswijk et al.
Our results are consistent with those of Grandi et al. and Schäfer et al., who reported the identification of a 90S RNP that contains 35 nonribosomal proteins and the U3 snoRNA (9, 33). They identified 27 SSU processome proteins that we had identified previously (Nop1, Noc4, Nop56, Mpp10, Imp3, Imp4, Sof1, Rrp5, Rrp9, Utp1, Utp2, Utp4, Utp5, Utp6, Utp7 Utp8, Utp9, Utp10, Utp11, Utp12, Utp13, Utp15, Utp16, Utp17, Utp18, Utp21, Utp22, and Emg1) but did not identify 2 other SSU processome components (Utp3 and Utp14). Due to the large overlap of protein components, it seems highly probable that we have independently characterized the same complex. Our results are also consistent with those of Krogan et al. and Peng et al., who used genomics and proteomic experiments to identify additional proteins required for ribosome biogenesis (17, 29).
One important difference between the interpretation of our results and those previously published is whether Enp1 is a component of the SSU processome-90S preribosome (3, 9, 33). We have shown here that tagged Enp1 does not detectably coimmunoprecipitate with other SSU processome components or the U3 snoRNA, compared to positive and negative controls, because Enp1 coimmunoprecipitated with the same amounts of Mpp10 and the U3 snoRNA as did the negative controls (Rpf2 and the untagged parental strain, YPH499). In addition, genetic depletion of Enp1 led to the accumulation of the 21S pre-rRNA, a phenotype different from that caused by the depletion of any other SSU processome component. The appearance of the 21S pre-rRNA upon Enp1 depletion suggests that it may participate in a later complex, as has been suggested by Milkereit et al. (25, 33). However, sedimentation of Enp1 on sucrose density gradients shows that Enp1 sediments at both 40S and 90S (33). These results therefore suggest that Enp1 may have multiple roles in ribosome biogenesis, and we cannot rule out the possibility that it may also be present in substoichiometric amounts in the SSU processome (3).
In addition to the SSU processome's role in pre-rRNA processing, it also likely has a role in RNA folding, as has been suggested previously (4, 39). Furthermore, the specific association of a subset of ribosomal proteins with the SSU processome suggests that the SSU processome is also an assembly intermediate for ribosome biogenesis. Therefore, folding of the pre-rRNA for processing is intertwined with ribosome assembly.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»