The Sequence of the 5′ End of the U8 Small Nucleolar RNA Is Critical for 5.8S and 28S rRNA Maturation

Abstract

Ribosome biogenesis in eucaryotes involves many small nucleolar ribonucleoprotein particles (snoRNP), a few of which are essential for processing pre-rRNA. Previously, U8 snoRNA was shown to play a critical role in pre-rRNA processing, being essential for accumulation of mature 28S and 5.8S rRNAs. Here, evidence which identifies a functional site of interaction on the U8 RNA is presented. RNAs with mutations, insertions, or deletions within the 5 -most 15 nucleotides of U8 do not function in pre-rRNA processing. In vivo competitions in Xenopus oocytes with 2 O-methyl oligoribonucleotides have confirmed this region as a functional site of a base-pairing interaction. Cross-species hybrid molecules of U8 RNA show that this region of the U8 snoRNP is necessary for processing of pre-rRNA but not sufficient to direct efficient cleavage of the pre-rRNA substrate; the structure or proteins comprising, or recruited by, the U8 snoRNP modulate the efficiency of cleavage. Intriguingly, these 15 nucleotides have the potential to base pair with the 5 end of 28S rRNA in a region where, in the mature ribosome, the 5 end of 28S interacts with the 3 end of 5.8S. The 28S-5.8S interaction is evolutionarily conserved and critical for pre-rRNA processing in Xenopus laevis. Taken together these data strongly suggest that the 5 end of U8 RNA has the potential to bind pre-rRNA and in so doing, may regulate or alter the pre-rRNA folding pathway. The rest of the U8 particle may then facilitate cleavage or recruitment of other factors which are essential for pre-rRNA processing.

Nucleolar localization elements in U8 snoRNA differ from sequences required for rRNA processing

“…Nonintronic snoRNAs, such as U8, which are transcribed from their own genes, are made with a monomethylguanosine cap that can be converted in the nucleus to a trimethylguanosine cap (Peculis & Steitz, 1994;Terns & Dahlberg, 1994;Cheng et al+, 1995;Terns et al+, 1995)+ Fluorescein-labeled U8 snoRNA that lacks a cap when injected is still localized in nucleoli, as can be seen in Figure 2+ As another control, U8 was synthesized with an A-cap rather than its usual G-cap (see Materials and Methods); U8 snoRNA with an A-cap cannot be trimethylated, although it is stable and can function in rRNA processing (Peculis & Steitz, 1994) and, as anticipated, localizes to nucleoli after 2 h (Fig+ 2)+ Therefore, the nucleolar localization in Xenopus oocytes does not depend on the presence or nature of the 59 cap on the injected U8 snoRNA+ U8 snoRNA synthesized with a monomethylguanosine cap was injected into oocytes for the studies described below, to emulate the in vivo situation as closely as possible+ Sequences in U8 snoRNA needed for rRNA processing are not required for nucleolar localization U8 snoRNA has been dissected by mutagenesis to define the sequences needed for rRNA processing (Peculis & Steitz, 1994;Peculis, 1997)+ Endogenous U8 snoRNA was depleted by antisense oligonucleotide injection into Xenopus oocytes, and mutated U8 snoRNA constructs were subsequently injected to assay if they could rescue the rRNA processing defect+ Those ex-periments indicated that the 59 region of U8 snoRNA was important for function in rRNA processing+ We have analyzed the mutated U8 snoRNAs used in those studies that were defective in rRNA processing to see if they can localize to nucleoli+ The failure to rescue rRNA processing could possibly reflect a secondary effect caused by inability of the mutant snoRNA to localize to nucleoli where rRNA processing occurs+ Alternatively, the sequences that were changed in the mutant snoRNA may be important only for rRNA processing, but not for nucleolar localization+ In fact, we will show here that the latter is the case for many sequences in U8+ Mutant U8 molecules examined for their ability to localize to nucleoli are listed in Figure 3+ Most of these are substitutions of sequence, but two mutants were constructed in this study that had regions of the molecule deleted (⌬82-107; ⌬St3,4)+ Table 1 summarizes previous (Peculis & Steitz, 1994) and present results of the ability of mutant U8 snoRNAs to rescue rRNA processing, and compares this to their nucleolar localization, as well as their association with the nucleolar protein, fibrillarin+ U8 snoRNA mutants that had substitutions of sequences found in loops and that lost function in rRNA processing were analyzed for their nucleolar localization+ As a positive control, we analyzed U8 loop mutant Lp4M, which is known to retain its function in rRNA processing (Peculis & Steitz, 1994) and therefore must presumably be localized in nucleoli+ Indeed, the Lp4M mutant localized to nucleoli (Fig+ 4)+ Similarly, loop mutants Lp1M and Lp2M localized to nucleoli (Fig+ 4), even though they cannot function in rRNA processing …”

“…U8 wild-type snoRNA localizes to nucleoli regardless of its cap structure+ Fluorescein-labeled U8 snoRNA with a G-cap, A-cap, or no cap at its 59 end was injected into Xenopus oocyte nuclei+ In some cases, the fluorescein label forms ring-like structures surrounding the rDNA core+ A lampbrush chromosome is present in the phase contrast view of the G-cap U8-injected material; it stains blue with DAPI, but is not labeled by fluorescein+ Other details as in Figure 1+ Bar is 10 mm+ resulting mutants are unstable 6 h or longer after injection, and their ability to function in rRNA processing cannot be determined (Peculis & Steitz, 1994)+ Nonetheless, they are sufficiently stable 2 h after injection that their localization can be assayed, and U8 snoRNA carrying substitutions in Box C or Box D failed to localize to nucleoli (Lange et al+, 1998)+ The same was true for U14 snoRNA, where even a single point mu-FIGURE 3. Mutations in U8 snoRNA+ The predicted secondary structure of Xenopus laevis U8 snoRNA is shown in the upper part of the figure, as in Peculis andSteitz (1993, 1994)+ Evolutionarily conserved Boxes C and D are indicated according to the consensus tabulated by Xia et al+ (1997)+ Regions containing substitutions (Peculis & Steitz, 1994) or deletions (this study) are enclosed by lines; the St2M mutation is a subset of the St2C substitution and is marked by a dotted line+ The sequence of wild-type U8 snoRNA and the mutants analyzed here are listed in the lower part of the figure+ A dot indicates that the nucleotide is the same as in wild-type U8 snoRNA; a dash indicates that the nucleotide is deleted+ tation in Box C or Box D prevented nucleolar localization (Lange et al+, 1998)+ Most of the snoRNAs in the Box C/D family have both ends of the molecule base paired together, and Boxes C and D flank either side of this terminal stem (Tycowski et al+, 1993)+ U3 and U8 snoRNAs are exceptions to this structural arrangement+ Nonetheless, it could be hypothesized that the two stems between Boxes C and D in U8 snoRNA act to hold the Boxes in the correct topographical arrangement relative to one another+ We have tested if alteration of this topography affects nucleolar localization by construction of two mutants where these stems have been altered+ In the ⌬82-107 mutant, one side of each stem has been deleted, thereby removing both stem structures (Fig+ 3)+ The 49 nt that remain between Boxes C and D in this mutant might fold differently or be unstructured compared with the wild type+ The more extensive deletion in the ⌬St3,4 mutant removes both stems, and only 4 nt remain between Boxes C and D+ When fluorescein-labeled transcripts of ⌬82-107 or ⌬St3,4 are injected into oocyte nuclei, both are localized to nucleoli (Fig+ 7)+ Therefore, sequences and structures of stems 3 and 4 that occur between Boxes C and D are not needed for nucleolar localization of U8 snoRNA+ These sequences and structures are not absolutely essential for rRNA processing either, as determined by a depletion and rescue experiment+ After the depletion of endogenous U8 snoRNA by antisense oligonucleotide injection into oocytes, the rRNA processing defects can be rescued by wild-type U8 snoRNA with or without fluorescein label (Fig+ 8)+ By comparison, the ⌬82-107 mutant gives almost complete rescue and the ⌬St3,4 mutant of U8 snoRNA gives partial rescue of rRNA processing+ Therefore, the third and fourth stems of U8 snoRNA are not important for nucleolar localization, nor do they play essential roles in rRNA processing+ …”

The structure of the ITS2-proximal stem is required for pre-rRNA processing in yeast

“…Previously one of us (B+A+P+) proposed that one role for the 59 end of U8 snoRNA in pre-rRNA processing in Xenopus oocytes was to transiently interact with prerRNA to facilitate formation of the ITS2-proximal stem between 5+8S and 28S rRNA (see Fig+ 2; Peculis, 1997b)+ A key feature of this model is that proper formation of the ITS2-proximal stem is a prerequisite for the processing events within ITS2 that lead to the accumulation of mature 5+8S and 28S rRNA (Peculis, 1997b)+ Here we directly tested this feature of the model using the well characterized genetic system available in yeast for studying rRNA processing (Nogi et al+, 1991a,b;Henry et al+, 1994)+ Two single, clustered mutations were made in which each of the 16 nt on one side of the ITS2-proximal stem was changed to disrupt the stem+ Both single mutations prevented accumulation of mature, large subunit rRNAs+ In contrast, a double, compensatory mutation that changed the primary sequence but preserved the base pairing potential of the stem did not prevent processing+ The double mutation also rescued a rRNA expression defect in strain NOY504, indicating it did not block other essential steps in ribosome biogenesis or translation+ These data demonstrate that the secondary structure of the ITS2-proximal stem is required for processing whereas the primary sequence is not essential for accurate processing, ribosome assembly, or translational function+ ITS2-proximal stem structure is required for processing accuracy, but its sequence affects efficiency…”

“…Other snoRNAs implicated in cleavage events are U3 and U14, two nearly ubiquitous snoRNAs that are required for accumulation of mature 18S rRNA (reviewed in Maxwell & Fournier, 1995;Tollervey & Kiss, 1997)+ U3 is essential for processing at a site within the 59 ETS (Hughes & Ares, 1991;Beltrame & Tollervey, 1992) and maintains the ability to base pair to a site within the 59 ETS despite phylogenetic sequence variation (Beltrame & Tollervey, 1992;Beltrame et al+, 1994)+ Depletion of U3, whether genetically in yeast (Beltrame & Tollervey, 1992;Beltrame et al+, 1994), or biochemically in Xenopus oocytes (Savino & Gerbi, 1991), results in an underaccumulation or complete absence of mature 17S/18S rRNA+ U14 has also been shown to be essential for cleavages that result in the accumulation of mature 17S/18S rRNA (Jarmolowski et al+, 1990;Li et al+, 1990)+ Other snoRNAs implicated in 17S/18S processing are less well characterized, having been identified in either vertebrates or yeast, but not both+ Depletion of U22 in Xenopus (Tycowski et al+, 1994), or of snR10 or snR30 in yeast (Tollervey, 1987;Girard et al+, 1992;Morrissey & Tollervey, 1993;Venema & Tollervey, 1996) affects accumulation of mature 17S/18S rRNA+ The mechanisms by which these snoRNAs function are not well defined+ U8 snoRNA has been shown to be essential for prerRNA processing in vertebrates (Peculis & Steitz, 1993)+ To date it is the only snoRNA shown to affect processing of the large subunit ribosomal RNAs+ Biochemical depletion of U8 snoRNA in Xenopus oocytes results in a lack of mature 5+8S and 28S rRNAs (Peculis & Steitz, 1993)+ Recently, a model for the role of U8 snoRNA in prerRNA processing in Xenopus was proposed (Peculis, 1997b)+ In this model, the 59 end of U8 base pairs to the 59 end of 28S in the pre-rRNA+ This interaction is proposed to be transient, serving to facilitate a subsequent, stable base pairing interaction between 28S and 5+8S rRNA sequences to form the ITS2-proximal stem+ The FIGURE 1. Pre-rRNA processing scheme in yeast+ Schematic of the genomic organization of rDNA and processing pathways of the pre-rRNA transcript+ The full-length precursor in yeast, 35S, is processed by alternative coexisting pathways, utilizing a variety of processing sites (denoted by gray letters and thin arrowheads on the precursor, 35S)+ Some of the first cleavages separate 18S rRNA from the 5+8S/25S+ Subsequent cleavages within ITS2 yield mature 5+8S and 25S and are shown by the heavy lines with arrows+ The sequence and structural requirements for these ITS2 processing events are focused upon in this ...…”

“…Pre-rRNA processing scheme in yeast+ Schematic of the genomic organization of rDNA and processing pathways of the pre-rRNA transcript+ The full-length precursor in yeast, 35S, is processed by alternative coexisting pathways, utilizing a variety of processing sites (denoted by gray letters and thin arrowheads on the precursor, 35S)+ Some of the first cleavages separate 18S rRNA from the 5+8S/25S+ Subsequent cleavages within ITS2 yield mature 5+8S and 25S and are shown by the heavy lines with arrows+ The sequence and structural requirements for these ITS2 processing events are focused upon in this work+ ITS: Internal Transcribed Spacer+ ITS2-proximal stem, a near perfect 15-base-pair helix in Xenopus, is the longest of the 3 base paired interactions between 5+8S and 28S+ This model for U8 snoRNA function and the role of the ITS2-proximal stem is based on four lines of evidence+ First, extensive mutagenesis studies and phylogenetic conservation of regions within the U8 sequence implicate the 59 end of U8 in critical sequence-dependent interactions (Peculis & Steitz, 1994;Peculis, 1997b)+ Second, 29-O-methyl oligoribonucleotides complementary to the 59 end of U8 RNA inhibit processing of the large subunit rRNAs in Xenopus oocytes (Peculis, 1997b)+ This inhibition is sequence specific, does not result in degradation of U8 RNP and can be rescued by subsequent injection of U8 snoRNA+ Thus, inhibition in this oligonucleotide competition experiment is presumably because of interference with base pairing of the 59 end of U8 RNA to its target+ Third, complementarity between the 59 end of U8 RNA and the 59 end of 28S rRNA is phylogenetically conserved, implicating this segment of 28S rRNA as a potential U8 target site (Peculis, 1997b)+ Fourth, 29-O-methyl oligoribonucleotides complementary to the 5+8S or 28S rRNA sequences in the ITS2-proximal stem inhibit processing of the large subunit rRNAs (Peculis, 1997b)+ While this fourth result is consistent with a requirement for the ITS2-proximal stem in processing, a direct test of this key feature of the model was lacking+ To provide such a test, we have taken advantage of the observation that this stable, base paired structure involving 28S and 5+8S rRNA is conserved in mature ribosomes across evolution (see Figs+ 2 and 6A)+ Thus, this interaction also occurs in yeast pre-rRNA, for which a genetic system to analyze mutations affecting rRNA processing is readily available (Nogi et al+, 1991a,b;Henry et al+, 1994)+ The results of our analyses in yeast demonstrate that the structure of the ITS2-proximal stem is essential for processing, whereas the sequence of the stem affects efficiency+…”