INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

18S, 5.8S, and 25-28S rRNAs are transcribed in the nucleolus of the eukaryotic cell by RNA polymerase I in a form of long precursor rRNA (pre-rRNA) that subsequently undergoes a number of processing cleavages to remove the external transcribed spacer sequences (ETS) and internal transcribed spacer sequences (ITS) and release the mature rRNAs. rRNA processing involves a number of small nucleolar RNAs (snoRNAs). U3 and U14 snoRNAs have been implicated in processing steps leading to 18S rRNA formation in both lower (U3 [20] and U14 [31, 32]) and higher (U3 [7, 24, 43] and U14 [27]) eukaryotes. In addition, 18S rRNA formation in vertebrates requires U22 (49), U17(E1), E2, and E3 snoRNAs (35), and in yeast it requires snR10 (48) and snR30 (36) snoRNAs. The role of snoRNAs in rRNA processing is distinct from the function of the majority of snoRNAs that serve as guide RNAs for rRNA modification (2'-O-ribose methylation and pseudouridine formation). In this paper, we investigate the role of U3 snoRNA in rRNA processing to form 18S rRNA.

U3 is the most abundant snoRNA in the cell and is essential for viability, as tested in yeast (20, 42). Vertebrate U3 snoRNA appears to determine the order of cleavage events leading to 18S rRNA formation (7). In Xenopus laevis, U3 snoRNA is required for cleavage at site 1 (5' end of 18S rRNA), site 2 (3' end of 18S rRNA), and site 3 (5' end of 5.8S RNA) (7, 43). In addition, U3 together with U14, U17(E1), and E3 snoRNAs are needed for cleavage near the 5' end of the ETS in many eukaryotes (12, 24). In yeast, U3 snoRNA is needed for cleavage at site A0 (near the 3' end of the ETS), site A1 (5' end of 18S rRNA), and site A2 (within the ITS1) (20).

What is the mechanism by which U3 snoRNA mediates these cleavages in pre-rRNA? It has been proposed that the single-stranded 5' hinge (4) and 3' hinge (8), which separate domains I and II of U3 snoRNA, base pair with the ETS, thus docking U3 snoRNA on the pre-rRNA substrate. These base-pairing interactions are supported by phylogenetic comparisons (8), and compensatory base changes validate the U3 5' hinge base pairing with the ETS of pre-rRNA in yeast (5). Once U3 snoRNA has docked on the pre-rRNA, it may base pair with sequences within the 18S pre-rRNA. In doing so, U3 snoRNA could act as a chaperone to prevent premature pseudoknot formation in 18S rRNA (19), similar to comparable interactions that occur in cis in bacteria (11). In yeast, three evolutionarily conserved regions in domain I of U3 snoRNA, the GAC element, box A', and box A (42), are complementary to sequences in 18S rRNA (19), and chemical modification supports the model that they base pair (34). Compensatory mutations validated the base pairing interaction between the 5' portion of box A and 18S rRNA in yeast but did not confirm the interaction between the 3' portion of box A and 18S rRNA (46). The interaction between box A' and 18S rRNA has not yet been experimentally tested in any organism. Base pairing between the GAC element and 18S rRNA can be drawn for yeast but not for vertebrates, calling its role into question. Therefore, further investigation is needed to explore the roles of the GAC element, box A', and box A in U3 snoRNA, and this is the focus of the present study.

In order to understand the mechanism for U3 snoRNA function in rRNA processing, we have created a systematic array of mutations in the 5' portion of the molecule, including the evolutionarily conserved GAC element, box A', and box A, and tested their effects in a functional assay. The results reported here have uncovered a new cleavage site in the 3' region of the ETS of vertebrate pre-rRNA, homologous with site A0 in yeast. In addition, by mutagenesis of U3 snoRNA we have uncoupled cleavage at sites 1 and 2, which normally occur in unison. Furthermore, specific nucleotides in U3 snoRNA that are required for distinct cleavages in pre-rRNA were identified: box A is essential for site A0 cleavage, the GAC-box and A' region is necessary for site 1 cleavage, and a few nucleotides at the 3' end of box A' are required for site 2 cleavage. Finally, based on these data, a new model is presented in which U3 snoRNA acts as a bridge to draw together the 5' and 3' ends of the 18S rRNA coding region within pre-rRNA.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Oligonucleotide injections for U3 snoRNA depletion and rRNA analysis. For U3 snoRNA depletion, stage 5 and 6 Xenopus laevis oocytes were injected with an oligonucleotide complementary to residues 39 to 54 of U3 snoRNA and rRNA was in vivo labeled by injection of ³²P-UTP as described earlier (7). Capped T7 RNA polymerase transcripts of wild-type or mutated U3 snoRNA were synthesized and injected into the depleted oocytes; subsequently, nuclear RNA was isolated as described previously (7). The RNA was electrophoresed on a denaturing 1.2% (wt/vol) agarose gel containing 6% (vol/vol) formaldehyde and electrotransferred onto a Nytran Plus membrane (Schleicher and Schuell, Keene, N.H.) for exposure to X-ray film (7). In some cases, the rRNA was analyzed by Northern blottings after radioactive rRNA on the filters had decayed for several months; filters were checked for the absence of residual radioactivity by using a Fuji X phosphorimager and BAS 1000 MacBas software. Northern hybridization was carried out in 6× SSC-1% (wt/vol) sodium dodecyl sulfate-5× Denhardt's solution -20% polyethylene glycol 6000 (wt/vol)-50 µg of sheared salmon sperm DNA/ml-100 µg of E. coli tRNA (Sigma, St. Louis, Mo.)/ml for 12 to 16 h at a temperature 5°C below the T_m of the hybrid and then washed as previously described (7) (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate, pH 7.0).

Hybridization probes. The following antisense oligonucleotides were used as specific probes against the X. laevis 5' ETS pre-rRNA: ETS-1 (nucleotides [nt] 30 to 47), 5'-CGGTCCTTTTTTCGGGCG-3'; ETS-2 (nt 475 to 498), 5'-GGCCTACTCTCCTTTCTTTCCGTC-3'; ETS-3 (nt 518 to 538), 5'-GGGGAGGGGGGGGGGAGGCGG-3'; ETS-4 (nt 549 to 572), 5'-GGGGGCCCCGCCCGGCCTGCCGCT-3'; ETS-5 (nt 627 to 650), 5'-CGGGTCGGCGCCCTGAGGCGTCAC-3'; ETS/18S (nt 700 to 715), 5'-GTAGCCACCTTTCCCG-3'.

U3 snoRNA mutagenesis. U3 snoRNA mutants were created by using the one-step PCR approach described previously (8), with the T7 promoter sequence at the 5' end of the 5' (sense) primer preceding 21 nt of U3 snoRNA sequence (mutated as shown in Fig. 1; the deletion primers were 21 nt long, and the insertion primers were 21 nt plus the length of the insertion). The U3 box A mutations were derived from a 5' (sense) primer with T7 promoter preceding 40 nt of X. laevis U3 snoRNA sequence (22) with the substitutions shown in Fig. 1. The U3 mutant sub 8-28 is the same as the box A+ mutant, and the U3 mutant sub 17-28 is the same as the box A mutant of Lange et al. (28). All deoxyoligonucleotides for U3 snoRNA mutagenesis were obtained from Genosys Biotechnologies (The Woodlands, Tex.) or Life Technologies GIBCO BRL (Gaithersburg, Md.). Oligo 39-54 for U3 snoRNA depletion was from Life Technologies GIBCO BRL. The following PCR cycle was used in the mutagenesis procedure: 94°C for 7 min, followed by 5 cycles of 94°C for 30 s, 48°C for 30 s, and 72°C for 30 s, and then 30 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 32 s, with the final extension at 72°C for 7 min. PCRs used the Amplitaq Gold PCR kit from Perkin-Elmer (Branchburg, N.J.) according to the protocol provided by the manufacturer.

View larger version (48K): [in this window] [in a new window]   FIG. 1.   U3 snoRNA mutagenesis. (Top) Boldface letters indicate nucleotide substitutions or insertions. The amount of rRNA is shown: ++, much; +, some; (+), little; , none. (Bottom) U3 nucleotides required for cleavage at site A0, 1, or 2 in pre-rRNA are bracketed.

snoRNA synthesis and purification. The PCR products were gel purified and cloned into the pT7 Blue (R) T-cloning vector (Novagen, Madison, Wis.), and all mutations were confirmed by sequencing. DNA of the plasmid constructs served as the template in subsequent PCRs to produce templates for in vitro T7 RNA polymerase transcription (see reference 28 for details). The PCR templates were purified with a QIAprep Spin Miniprep Kit (Qiagen, Valencia, Calif.). The RNA concentration was estimated by comparison to reference samples with known concentrations on 8% (wt/vol) acrylamide-7 M urea gels electrophoresed in TBE (90 mM Tris-borate, 2 mM EDTA, pH 8.0) and stained with methylene blue. The concentration was confirmed by spectrophotometry (A₂₆₀). The concentration of snoRNA used for oocyte injection in rescue experiments spanned a range of 0.125 to 1.5 ng/nl in order to find the optimum for rescue.

Primer extension. Primer extension was performed as follows. The annealing mixture contained 3 µl of total RNA from 10 to 20 oocyte nuclei, 1 µl of 5× First Strand buffer (250 mM Tris-HCl, pH 8.3, 375 mM KCl, 15 mM MgCl₂), and 0.5 to 1 pmol of ³²P-end-labeled primer. The primer was either the ETS/18S oligonucleotide used for Northern blottings or an 18S tag (5' CGT CAC ACT CGA GGG CGA TCG 3') complementary to nt 277 to 297 from the 5' end of Xenopus 18S rRNA. The nucleotides in bold were substitutions in the tag sequence compared to the wild type, in order to distinguish nascent 18S rRNA from preexisting 18S rRNA. Samples were boiled for 5 min, slowly cooled for ~30 min, and transferred to a 42°C water bath. Then, 30 µl was added, which contained 7 µl of 5× First Strand buffer, 2 µl of 0.1 M dithiothreitol, 2 µl of deoxynucleotide triphosphate mix (10 mM concentrations of each; Clontech, Palo Alto, Calif.), 18.5 µl of H₂O, and 100 to 200 U of SuperScript reverse transcriptase (Life Technologies GIBCO BRL, Gaithersburg, Md.). The reaction mixture was incubated for 60 min and stopped by a chloroform extraction and ethanol precipitation. Labeled products of the primer extension reaction were resolved on a 6% denaturing acrylamide gel next to a sequencing ladder prepared by dideoxy sequencing with the same end-labeled primer of a plasmid construct containing Xenopus pre-rRNA.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Mapping cis-acting elements in the GAC-box A' region of U3 snoRNA. The 5' end of U3 snoRNA sequence has a high degree of phylogenetic conservation, containing the GAC element, box A', and box A, listed in order from the 5' end (42). In order to find out if these conserved regions of U3 snoRNA reflect areas of functional importance for rRNA processing, we created a series of U3 mutations (Fig. 1) and systematically tested them for the ability to promote mature 18S rRNA formation by an in vivo U3 depletion-rescue assay in Xenopus oocytes. In this assay, endogenous intact U3 snoRNA is destroyed by injection of an antisense oligonucleotide into Xenopus oocyte nuclei (7, 43). Subsequently, wild-type or mutant U3 snoRNA is injected together with ³²P-UTP to analyze rRNA processing by gel electrophoresis. In the absence of U3 snoRNA disruption, the 40S rRNA precursor, processing intermediates, and the mature 18S and 28S rRNA species are seen (Fig. 2). After destruction of intact U3 snoRNA, 18S rRNA was not produced (Fig. 2), but its production was restored by injection of wild-type U3 snoRNA (Fig. 2). We investigated whether mutant U3 snoRNA could also restore 18S rRNA formation. The failure to make 18S rRNA was not due to instability of the injected U3 snoRNA mutant, because a ³²P-labeled mutant carrying a large sequence substitution spanning boxes A' and A remained stable for 2 h after oocyte injection (28). Also, this and other U3 transcripts mutated at the 5' end were stable for 18 h after injection (data not shown). Moreover, the 5' portion of U3 snoRNA is not essential for nucleolar localization of U3 snoRNA (28), indicating that all the mutations studied here can localize to the nucleolar compartment where pre-rRNA processing takes place.

View larger version (66K): [in this window] [in a new window]   FIG. 2.   Mutations in the 5' portion of Xenopus U3 snoRNA affect rRNA processing. Endogenous U3 snoRNA was not depleted () or was depleted (+) by injection into Xenopus oocytes of an antisense oligonucleotide complementary to nt 39 to 54. Subsequently, either wild-type (WT) or mutant (Fig. 1) synthetic U3 was injected into the oocytes for rescue. In vivo labeling was done by subsequent injection of 32P-UTP to trace changes in rRNA processing after U3 snoRNA depletion and rescue. The sizes of nuclear pre-rRNA and rRNA are indicated; note the novel 19S and 18.5S pre-rRNAs. Results of injection of additional U3 mutants not shown here are summarized in Fig. 1.

Functional importance of box A of U3 snoRNA. Mutation of U3 box A (sub 17-28) prevented production of 18S rRNA, and neither 19S nor 18.5S pre-rRNA appeared (Fig. 3A, lane 6). The same was also true for a U3 snoRNA mutation carrying a substitution at the 5' end of box A (sub 17-19) or the 3' end of box A (sub 23-28) where no 18S, 18.5S or 19S pre-rRNA accumulated (Fig. 3A, lanes 2 and 5, respectively), indicating that the left and right portions of U3 box A seem to have the same function. This is in contrast to box A', where mutation of the right but not the left side produced 18.5S pre-rRNA. Nt 22 in the middle of box A is highly exposed (34), but its substitution (sub 22) or deletion ( 22) had no adverse effect on the function of box A (Fig. 1; data not shown). Box A had a more severe effect when mutated than was true for the GAC-box A' element, since neither 19S nor 18.5S were formed. Moreover, mutations in U3 snoRNA spanning box A' through box A (U3 sub 8-28) prevented formation of 18S, 18.5S, and 19S pre-rRNA (Fig. 3A, lane 8), showing that the effects of box A mutation are dominant over those of box A' and its 3' flanking nucleotides where 19S and 18.5S pre-rRNAs were produced.

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Effects on rRNA processing of substitutions in U3 box A (A), deletions (B), or insertions between box A' and box A (C). Note that endogenous U3 snoRNA was not depleted in panel B, lanes 1 to 3. Other details are as described for Fig. 2. WT, wild type.

Spacing between box A and the 5' end of U3 snoRNA is critical. Mutations upstream of U3 box A caused 19S pre-rRNA to appear, but no 19S pre-rRNA was formed after mutation of U3 box A itself. Thus, nt 15 and 16 appear to be a boundary between different functional elements of the 5' region of U3 snoRNA. We tested whether the spacing between U3 box A and the GAC-box A' element was important for function in rRNA processing. Deletion of nt 15 and 16 in U3 snoRNA inhibited the formation of 18S rRNA, and 18.5S pre-rRNA appeared instead (Fig. 3B, lane 2). This differs from substitution of the same 2 nt in U3 (sub 15-16), where some 18S rRNA was made and 18.5S was absent (Fig. 3B, lane 4). In contrast, deletion of U3 nt 1 to 6 yielded the same result as substitution of these nucleotides (Fig. 2, lanes 4 and 7, respectively). Thus, a distinct spacing between box A and the GAC-box A' region appears to be necessary for proper U3 snoRNA function in rRNA processing. This notion was further substantiated by examination of a series of insertions between nt 16 and nt 17 (directly at the left border of U3 box A). Insertions of 4, 7, or 8 nt at this position in U3 snoRNA prevented any significant formation of 18S rRNA and instead 19S pre-rRNA strongly accumulated (Fig. 3C, lanes 3 and 4, and data not shown). This result was independent of the sequence used for insertion, and two different sequences used for the 4-nt insertion yielded the same effect (Fig. 3C, lanes 3 and 4).

Identification of 19S and 18.5S rRNA processing intermediates. Northern blottings were performed with probes for several locations in pre-rRNA to identify 19S and 18.5S pre-rRNAs. RNA that was used for Northern blots (Fig. 4 and 5, lanes N) was compared to RNA from untreated oocytes that was ³²P-labeled in vivo (Fig. 4 and 5, lanes MW) or RNA from U3-depleted oocytes injected with mutated U3 snoRNA to produce 18.5S pre-rRNA (Fig. 4) or 19S pre-rRNA (Fig. 5). Both 18.5S and 19S pre-rRNA were detected on Northern blots with probes for the 18S coding region (Fig. 4, lane 11, and Fig. 5, lane 10, respectively), suggesting that these novel species are precursors to 18S rRNA. As expected, the 18S coding region probes also detected 18S rRNA as well as its 40S and 20S precursors in these Northern blots.

View larger version (46K): [in this window] [in a new window]   FIG. 4.   Identification of the novel 18.5S pre-rRNA by Northern blottings. Samples from in vivo labeling experiments (as described for Fig. 2) that showed strong accumulation of the novel 18.5S pre-rRNA (produced by rescue with 15-16 mutant U3) or 19S pre-rRNA (in lanes 6 and 8 and produced by rescue with 1-6 U3 mutant snoRNA) were resolved on denaturing agarose gels, electrotransferred onto Nytran Plus membranes, and exposed to X-ray film (lanes 32P). The filters were allowed to decay for up to 1 year and were hybridized with the radioactive probes shown by the bars under the map, derived from the ETS, 18S coding region, or ITS1. Lanes N, Northern blots of the 32P lane in the same panel after its radioactivity had decayed; the same lanes are linked by a bracket. Probes 18S-1 through ITS1-4 detected 18.5S pre-rRNA (+ above map of pre-rRNA). Lanes MW, in vivo-labeled pre-rRNA and rRNA from unperturbed oocytes as molecular weight markers.

View larger version (45K): [in this window] [in a new window]   FIG. 5.   Identification of the novel 19S pre-rRNA by Northern blottings. Samples from in vivo labeling experiments (as described for Fig. 2) that showed strong accumulation of the novel 19S pre-rRNA (produced by rescue with 5' cb, 5' ncb, or 1-6 U3 mutant snoRNA) were allowed to decay and hybridized with probes from the ETS, 18S coding region, or ITS1 (see Fig. 4 for details). Note that 19S pre-rRNA was detected by probes from ETS-3 to ITS1-4 (+ above map of pre-rRNA).

View larger version (26K): [in this window] [in a new window]   FIG. 6.   Summary of cleavages that produce 20S, 19S, and 18.5S pre-rRNA. Results of Northern blots shown in Fig. 4 and 5 and data not shown are summarized here. Xenopus 20S, 19S, and 18.5S pre-rRNA all share the same 3' end (produced by cleavage at site 3 in the ITS1), but their 5' ends vary as indicated. The 5' end of 20S pre-rRNA is the same as the 5' end of 40S pre-rRNA (26, 44). X indicates the inhibited cleavage sites, resulting in accumulation of 20S, 19S, or 18.5S pre-rRNAs. Cleavage sites in yeast pre-rRNA are indicated (reviewed in reference 52) for comparison to sites in vertebrate pre-rRNA.

View larger version (26K): [in this window] [in a new window]   FIG. 7.   Mapping the novel A0 cleavage site in pre-rRNA by primer extension. RNA extracted from nuclei of unperturbed Xenopus oocytes (lane 1), oocytes depleted of endogenous U3 snoRNA (lane 2), or U3 depleted or rescued with U3 snoRNA mutants that promote strong accumulation of 19S pre-rRNA (lane 3) or 18.5S pre-rRNA (lane 4) was used as the template for primer extension; the 32P-labeled products of reverse transcription shown here next to sequencing lanes primed with the same primer were resolved on a 6% acrylamide denaturing gel. Site A0 cleavage was detected in pre-rRNA using a primer complementary to the 3' end of the ETS (open box with arrow; same as ETS/18S probe in Fig. 4 to 6); site 1 cleavage was specifically detected in plasmid-expressing tag containing 18S rRNA using a primer complementary to the tag region near the 5' end of 18S rRNA (open box with arrow).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Regions in vertebrate U3 snoRNA needed to form 18S rRNA. The conclusions from the dissection of Xenopus U3 snoRNA are summarized in Fig. 1. We found that nt 4 to 14 are required for site 1 cleavage of pre-rRNA, as their mutation obliterates this cleavage and 19S pre-rRNA accumulates. This is the first time in any organism that a function has been shown for the conserved GAC-box A' element, which seems to be a functional continuum. Within this region, nt 11 to 13 appear to have an additional role, being important also for site 2 cleavage in pre-rRNA (18.5S pre-rRNA accumulates). This is the first time that site 2 cleavage has been uncoupled from site 1 cleavage by mutation of U3 snoRNA. Finally, our data show that Xenopus U3 box A is required for the newly discovered site A0 cleavage, and mutation of this area of U3 snoRNA blocks any further processing of 20S pre-rRNA, which is a precursor to 18S rRNA.

View larger version (25K): [in this window] [in a new window]   FIG. 8.   Model for vertebrate U3 snoRNA interactions with pre-rRNA. The ETS of Xenopus pre-rRNA is drawn in dark blue and the 18S rRNA coding region in light blue; U3-dependent cleavage sites A0, 1, and 2 are in green. Xenopus U3 snoRNA is in red. Domain II of U3 is required for nucleolar localization (28) and for site 3 cleavage (7). The 5' hinge and 3' hinge of U3 snoRNA are proposed to dock U3 on pre-rRNA by base pairing with sequences in the ETS (8). Putative base-pairing interactions between the 5' region of U3 snoRNA and sequences in the ETS and 18S coding region of pre-rRNA are indicated (see text). U3 snoRNA is proposed to act as a bridge to draw together cleavage sites A0, 1, and 2. Solid lines indicate base pairs between U3 snoRNA and pre-rRNA confirmed by compensatory base changes in yeast (5, 46), and dotted lines denote putative base pairs deduced by sequence complementarity and phylogenetic comparisons. Alternative base pairing interactions between U3 box A' and the 5' or 3' end of the 18S coding region are shown. *, nucleotides in 18S rRNA that base pair to form the central pseudoknot (19). A computer-based (M-fold 3.0) model is shown for the ETS between sites A0 and 1, but bars for base pairing are not included. Pseudouridine () has been mapped to nt 8 and 12 of vertebrate U3 snoRNA (14, 41), though not studied directly in Xenopus.

U3-dependent cleavage at site A0. We suggest that box A helps to position U3 snoRNA properly on pre-rRNA to allow U3-dependent cleavage at site A0 to occur. We found that when box A is mutated in Xenopus U3 snoRNA, there is inhibition of cleavage at site A0 and at sites 1 and 2 posited to occur later in the processing pathway. It has been hypothesized that the 3' portion of box A base pairs with sequences in the 18S coding region (nucleotides ₁₁₅₈-AGGAA-₁₁₆₂ in Xenopus) to prevent pairing with nt 12 to 15 of 18S, thus blocking premature pseudoknot formation (19); the nucleotides that base pair to form the pseudoknot in mature 18S rRNA are indicated in Fig. 8. However, substitution of this AGGAA sequence in yeast 18S rRNA and the compensatory mutation in the complementary region in U3 box A did not restore 18S rRNA formation, thus calling into question the importance of its putative base pairing with U3 box A (46). As an alternative, we note that the same region of box A in U3 could pair with nucleotides ₄₈₂-AGAAA-₄₈₆ in the Xenopus ETS, preceding cleavage site A0 (Fig. 8). Moreover, mutation of these nucleotides in Xenopus U3 snoRNA (sub 23-28; Fig. 3) inhibits cleavage at site A0. Site A0 has not yet been mapped in most organisms, but in Trypanosoma brucei where its position is known (17), it is also 8 nt downstream from 5 nt in the ETS that have the potential to base pair with the same 3' portion of U3 box A (data not shown; ETS sequence from D. Campbell and T. Hartshorne, personal communication).

U3-dependent cleavage at site 1 in pre-rRNA. The 5' end of box A has been proven by compensatory base changes to base pair with sequences near the 5' end of 18S rRNA (46), located in the loop of a stem-loop structure shown to be a determinant of cleavage at site A1 in yeast (47, 53). This interaction between the 5' end of U3 box A and 18S rRNA is universal and can also be drawn for higher organisms (19) (Fig. 8 for Xenopus); the base pairing favorably positions vertebrate U3 snoRNA close to the cleavage sites A0 and 1. Since mutation of Xenopus box A inhibits cleavage at site A0, the potential importance of the 5' end of box A for site 1 cleavage, which is later in the processing pathway, was not revealed.

Role of U3 box A' for cleavage at sites 1 and 2 in pre-rRNA. We have shown here that residues in Xenopus U3 box A' are required for cleavage at site 1 and at site 2 of pre-rRNA. As shown in Fig. 8, box A' potentially base pairs with sequences near the 5' end of the 18S coding region (19), juxtaposing it close to cleavage site 1. This potential interaction is evolutionarily conserved (19). We propose that subsequently U3 box A' loses its pairing interaction with the 5' end of 18S rRNA and instead it and adjacent nucleotides enter into new base pairing with sequences near the 3' end of 18S rRNA in the pre-rRNA (Fig. 8). This would have the effect of drawing the 5' and 3' ends of 18S rRNA close together so that cleavage at these two sites could be coordinated. In essence, U3 snoRNA may act as a bridge between the two ends of 18S rRNA, as has been hypothesized previously that a snoRNA might replace the base-paired stem (37) that flanks 16S-18S rRNA in bacteria (54) and yeast. The bases involved in this putative interaction are evolutionarily conserved and identical in 18S rRNA and U3 snoRNA of all vertebrates. We have shown here that mutation of nt 11 to 13, which disrupts the putative base pairing with the 3' end of 18S rRNA, inhibits cleavage at site 2 but not at site 1, giving rise to 18.5S pre-rRNA (Fig. 2). Presumably, in these cases, the 3' end of 18S rRNA in the pre-rRNA cannot be drawn close to its 5' end nor is it associated with U3 snoRNA, and site 2 cleavage fails to occur. It appears that cleavage at site 1 is a prerequisite for cleavage at site 2, since cleavage at site 2 has never been observed by anyone without cleavage at site 1.

Differences in rRNA processing between vertebrates and yeast. While there are several similarities in rRNA processing between vertebrates and yeast, there are also significant differences. Unlike vertebrates, formation of the true 3' end of yeast 18S rRNA (site D) seems to occur in the cytoplasm (50) and therefore appears to be independent of U3 snoRNA. However, at an earlier step in yeast rRNA processing, U3-dependent cleavage occurs somewhat downstream in the ITS1 at site A2 (Fig. 6). As discussed above, it remains unknown if Xenopus site 3' is a composite of yeast sites A2 and A3.