Down-regulation of RNA Helicase II/Gu Results in the Depletion of 18 and 28 S rRNAs in Xenopus Oocyte*

Genetic manipulations have revealed the functions of RNA helicases in ribosomal RNA (rRNA) biogenesis in yeast. However, no report shows the role of an RNA helicase in rRNA formation in higher eukaryotes. This study reports the functional characterization of the frog homologue of nucleolar RNA helicase II/Gu (xGu or DDX21). Down-regulation of xGu in Xenopus laevis oocyte using an antisense oligodeoxynucleotide results in the depletion of 18 and 28 S rRNAs. The disappearance of 18 S rRNA is accompanied by an accumulation of 20 S, indicating that xGu is critical in the processing of 20 to 18 S rRNA. The degradation of 28 S rRNA into fragments smaller than 18 S is also associated with a specific decrease in the level of xGu protein. These effects are reversed in the presence of in vitro synthesized wild type xGu mRNA but not its helicase-deficient mutant form. Similar aberrant rRNA processing is observed when antibody against xGu is microinjected. The involvement of xGu in processing of rRNA is consistent with the localization of Gu protein to the granular and dense fibrillar components of PtK2 cell nucleoli by immunoelectron microscopy. Our results show that xGu is involved in the processing of 20 to 18 S rRNA and contributes to the stability of 28 S rRNA in Xenopus oocytes.

Genetic manipulations have revealed the functions of RNA helicases in ribosomal RNA (rRNA) biogenesis in yeast. However, no report shows the role of an RNA helicase in rRNA formation in higher eukaryotes. This study reports the functional characterization of the frog homologue of nucleolar RNA helicase II/Gu (xGu or DDX21). Down-regulation of xGu in Xenopus laevis oocyte using an antisense oligodeoxynucleotide results in the depletion of 18 and 28 S rRNAs. The disappearance of 18 S rRNA is accompanied by an accumulation of 20 S, indicating that xGu is critical in the processing of 20 to 18 S rRNA. The degradation of 28 S rRNA into fragments smaller than 18 S is also associated with a specific decrease in the level of xGu protein. These effects are reversed in the presence of in vitro synthesized wild type xGu mRNA but not its helicase-deficient mutant form. Similar aberrant rRNA processing is observed when antibody against xGu is microinjected. The involvement of xGu in processing of rRNA is consistent with the localization of Gu protein to the granular and dense fibrillar components of PtK2 cell nucleoli by immunoelectron microscopy. Our results show that xGu is involved in the processing of 20 to 18 S rRNA and contributes to the stability of 28 S rRNA in Xenopus oocytes.
Nucleoli are distinct phase-dense organelles inside the nucleus, the major function of which is to synthesize, process, and package ribosomal RNA (1). Additional functions have recently been assigned to nucleoli including production of signal recognition particles and transient sequestration of proteins involved in the regulation of gene transcription, cell cycle progression, and stem cell proliferation (2)(3)(4)(5). The recent identification of ϳ350 additional nucleolar proteins may eventually lead to additional functions for the nucleolus (6,7).
Electron microscopic examination of a nucleolus shows three structurally well defined regions: the fibrillar center (FC), 1 the dense fibrillar component (DFC), and the granular component (GC). Transcription of rRNA genes is thought to occur in the DFC and/or at the FC/DFC borders (8 -10), and early processing of the newly synthesized mammalian 47 S primary precursor transcript occurs within the DFC (11,12). RNA polymerase I catalyzes the synthesis of 47 S pre-rRNA, which is then processed into the mature 18, 5.8, and 28 S rRNAs. This processing resembles an assembly line involving numerous proteins and small nucleolar RNAs. Ribose methylation and pseudouridilation of the 47 S precursor RNA, guided by small nucleolar RNAs, occur prior to nucleolytic cleavages (13,14). Final processing and assembly of the rRNAs, together with ϳ80 ribosomal proteins, occur within the GC region and result in the formation of preribosomes that are eventually exported to the cytoplasm to become major components of the translation machinery (15,16).
Although the general processing pathway is conserved through evolution, the mechanism of rRNA production tends to be more complex when going from yeast to human. The process is better defined in yeast because of available genetic screenings (17,18). Most yeast factors have mammalian counterparts but because a mammalian nucleolus is more complex in structure and function, not all nucleolar proteins in higher eukaryotes have yeast homologues. Furthermore, if a mammalian homologue exists, it cannot always reverse the mutant yeast phenotype. For example, mammalian nucleolin, a homologue of the yeast NSR1, cannot complement the yeast nsr1-⌬ mutant (19). This shortcoming in the yeast system dictates that a different system must be employed to identify protein function.
The Xenopus laevis oocyte offers an alternative system for functional analyses of proteins with advantages that include large size and convenient physical and biochemical manipulations. Factors that may alter gene expression are conveniently microinjected, and the oocyte fractions can then be analyzed. This technique has been used to study proteins and RNAs involved in rRNA processing in the X. laevis oocyte (20 -26). Despite the convenience of using frog oocytes, details on the mechanism of rRNA synthesis in higher eukaryotes lag behind when compared with those in yeast. At least 16 putative helicases or DEAD box proteins have been genetically and biochemically implicated in yeast rRNA production (27), but none has been demonstrated in higher eukaryotes.
A nucleolar protein hypothesized to function in the biogenesis of rRNA is RNA helicase II/Gu (Gu). We previously reported the cloning and characterization of mammalian Gu, a protein with 5Ј to 3Ј RNA unwinding activity (28,29). Gu also contains a functionally distinct domain at its C terminus that is able to introduce secondary structure to single-stranded RNA (30). We recently identified Gu␤, a paralogue of the original Gu, which we now call Gu␣. The two proteins are products of structurally related adjacent genes on human chromosome 10 and may have evolved by gene duplication (31). The shuttling of Gu␣ between the nucleolus and nucleoplasm (32) supports our recent report on the activation of c-Jun-regulated transcriptions by Gu␣ (33). The localization of Gu␣ in the nucleolus and its RNA unwinding and annealing activities indicate a possible role in the production of highly structured rRNAs.
To gain more information about the role of Gu␣ in rRNA production, we cloned its Xenopus homologue. In this paper, we demonstrate that down-regulation of Gu␣ results in aberrant rRNA production in Xenopus oocytes. This is the first report showing the function of an RNA helicase in the biogenesis of rRNA in higher eukaryotes.

EXPERIMENTAL PROCEDURES
Cloning of the X. laevis Gu␣ Homologue-Using the mouse Gu␣ cDNA (34), the EST Data Bank was searched for its frog homologue. This search resulted in two almost identical X. laevis sequences; both are homologous to the 3Ј two-thirds of the mouse Gu␣ cDNA sequence. The sequences were designated xGu-1 (GenBank TM accession number AF302423) and xGu-2 (GenBank TM accession number AF302422).
To obtain full-length cDNA clones, a X. laevis oocyte cDNA library (Clontech) was screened according to standard protocols. A probe was prepared based on the obtained EST Data Bank sequences. BV602 (5Ј-ACCTGGCCGTGTGAGAGATCTTGTCCA-3Ј) and BV603 (5Ј-GAT-GAACATAAGCATCTGCTTCCTTTG-3Ј) were used to synthesize partial xGu cDNA from a liver total RNA using the One-Step RT-PCR kit (Qiagen). The 600-nucleotide cDNA fragment was randomly labeled (50 ng) with [␣-32 P]dCTP using a RadPrime DNA labeling system (Invitrogen). Prehybridization and hybridization of the membranes with the labeled probe were carried out according to standard procedures (35). Positive plaques were purified and rescreened according to the procedure that came with the cDNA library.
Lambda phage clones were converted to phagemid using BM25.8 Escherichia coli cells. Phagemid DNA was purified and analyzed using EcoRI and XbaI. The clones with the longest inserts were sequenced. This screening resulted in the isolation of two clones with full-length xGu-1 and xGu-2 sequences.
Isolation of Total RNA from Xenopus Tissues-Oocyte-positive human chorionic gonadotropin-tested female X. laevis were purchased from Xenopus Express. Different tissues were dissected from the frogs and stored at Ϫ80°C. TRIZOL reagent (Invitrogen) was used to isolate total RNA from 100 mg of homogenized tissue. RNA samples were quantified spectrophotometrically, and the quality was determined by formaldehyde gel electrophoresis.
Northern Blot Hybridization-A mixture of total RNA (10 g in 5 l of H 2 O) was treated according to standard protocol and separated on a 1.4% agarose-formaldehyde gel, blotted onto nitrocellulose membrane, and analyzed by Northern blot (35) using the same 32 P-labeled probes. The analysis of xGu mRNA was done with the same 32 P-labeled probe described above. The analysis of rRNAs used 32 P end-labeled oligodeoxynucleotides such as 18 S-5Ј (5Ј-TTGAGACAAGCATATGCTACT-3Ј) and 28 S-5Ј (5Ј-GTCGCCGCGTCTGATCTGAGG-3Ј). The level of 20 S rRNA was analyzed using a mixture of 32 P end-labeled ITS1A (5Ј-CG-AGACCCCCCTCACCCGGAGAGAGGGAAGGCGCCCGCCGCACCCT-CCCCGCGG-3Ј) and ITS1B (5Ј-GCGGGCGTCTCTCTCTCTCTCCGC-GGGGAGGGTGCGGCGGG-3Ј) as described previously (36).
Western Blot Analysis-Large scale preparation of oocyte extract was carried out by freon extraction to remove yolk protein as described (37). Approximately 5 g of this extract was analyzed by a colorimetric Western blot method using alkaline phosphatase.
For small scale preparations, germinal vesicles were manually isolated with forceps and boiled in Laemmli buffer. The samples were loaded onto 10% polyacrylamide-SDS gels and analyzed either by a colorimetric method or by a chemiluminescence detection system using an ECL Plus kit (Amersham Biosciences).
RT-PCR-A One-Step RT-PCR kit was used to quantitatively analyze mRNA levels using the sets of primers listed in Table I. Each RT-PCR contained 100 ng of total RNA, 5 l of 5ϫ RT-PCR buffer, 5 l of 5ϫ Q-solution, 200 M dATP, 200 M dTTP, 200 M dGTP, 200 M dCTP, 0.5 l of 10 units/l RNAsin, 1 l of RT-PCR enzyme mix, and 2 M each primer in 25 l of total volume. The RT-PCR was carried out at 50°C for 30 min and at 95°C for 15 min. Amplification was done at 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min (30 cycles). The reaction was extended at 72°C for 10 min. The RT-PCR products were analyzed in a 1% agarose gel, visualized under UV light, and photo-graphed using the Eagle Eye II system (Stratagene). The intensity of the DNA bands was measured in pixels with the Eagle Eye II system. ␤-Actin bands were used as internal controls.
Cellular Localization of xGu-A X. laevis kidney cell line (A6) was obtained from ATCC. The cells were grown on a chamber slide containing 75% National Cancer Tissue Culture 109 medium (Invitrogen), 15% water, and 10% fetal bovine serum in a 5% CO 2 incubator at 26°C. The cells were analyzed by double indirect immunofluorescence (38) using anti-xGu and anti-xC23 antibodies. The antibody against X. laevis nucleolin/C23 (b6 -6E7) was obtained from the Developmental Studies Hybridoma Bank at the University of Iowa.
Site-directed mutation of the DEVD motif to ASVD was carried out by PCR. The full-length mutant was also subcloned into the pGEX 4T-3 vector. All of the fusion proteins were expressed and purified described previously for GST-human Gu␣ (28).
For RNA helicase assays, a partially double-stranded RNA with a 34-nucleotide duplex was synthesized from pBluescript vector as reported previously (Ref. 28; see also Fig. 8). Annealing of RNAs, gel purification of the substrates, and the helicase assays were done as described (28,29) with minor modifications. Briefly, RNA helicase assays were carried out in a 20-l reaction mixture containing 20 mM HEPES-KOH, pH 7.6, 2 mM dithiothreitol, 3 mM MgCl 2 , 3 mM ATP, 0.1 M KCl, 2 units of RNAsin, 50 fmol of 32 P-labeled double-stranded RNA substrate, and the enzyme fraction at 37°C for 30 min. The reactions were stopped by adding 5 l of stop buffer (0.1 M Tris-HCl, pH 7.4, 20 mM EDTA, 1.3% SDS, 0.1% Nonidet P-40, 0.1% bromphenol blue, 0.1% xylene cyanol, 42% glycerol, and 0.8 mg/ml proteinase K) to the 20-l reaction mix, and incubated at 37°C for 5 min prior to loading of 12.5 l of reaction mix onto 10% SDS-PAGE gel. After electrophoresis at 100 volts, the gel was dried and analyzed by phosphorus imaging.
Assays for RNA folding activity were done as described previously (30). An [␣-32 P]GTP-labeled T7 RNA polymerase transcript from a HindIII-cut pBluescript II KS plasmid was synthesized, gel-purified, and dissolved in buffer containing 20 mM HEPES, pH 7.6, 0.2 M KCl, 0.1 mM EDTA. The single-stranded RNA substrate was appropriately diluted, boiled for 10 min, and cooled on ice just before using. RNA folding assay was done in 20 l of reaction buffer containing 20 mM HEPES-  Production of Anti-xGu Antibody-The cDNA region of xGu-1 that encodes amino acids 245-759 (92% identical to xGu-2) was amplified by RT-PCR using BV609 (5Ј-AAGCTAAATGCTAGCCAACAGCCATTG-GCTAGAGG-3Ј) and BV610 (5Ј-TGTTCGGACCTCGAGACGGCCC-CT- TCTAAACCCC-3Ј) and 1 g of frog liver total RNA using the Super-Script One-Step RT-PCR System (Invitrogen). The amplified cDNA fragment was digested with NheI and XhoI and subcloned into the pTYB3 expression vector (New England Biolabs, Inc.).
Protein expression was done by isolating plasmid DNA from the DH5␣ clone and transforming BL21-CodonPlus competent cells (Stratagene). The recombinant protein was purified according to the IMPACT T7: One-Step Protein Purification System protocol (New England Biolabs, Inc.). The purified protein was equilibrated overnight against a dialysis buffer (20 mM HEPES, pH 7.6, 0.1 mM EDTA, 0.5 M NaCl, 0.1% Triton X-100) and sent to Rockland, Inc. for antibody production in rabbit.
Affinity Purification of Anti-xGu Antibody-A polyclonal antibody was raised against xGu-1 protein as described above. Affinity purification was carried out using xGu-2 protein to purify antibody that recognized both xGu-1 and xGu-2. GST and GST-xGu-2 were separately immobilized onto cyanogen bromide-activated Sepharose (Sigma). The rabbit serum containing anti-xGu antibody was diluted 3-fold in 1ϫ PBS and passed twice through a GST affinity column (1 ml) to eliminate nonspecific binding. The flow-through from this GST column was passed through a GST-xGu-2 affinity column (1 ml) twice. The GST-xGu-2 column was washed consecutively with 15 ml of 1ϫ PBS, 10 ml of 1ϫ PBS containing 1.8 M KCl (pH 7.6), and 10 ml of 1ϫ PBS or until absorbance of the wash at 280 nm was 0. Bound antibodies were eluted with 0.2 M glycine (pH 2.2), and each 1-ml fraction was collected into a tube containing 0.4 ml of 1 M Tris-HCl (pH 8.0). All of the fractions were kept on ice, and absorbance at 280 nm was measured. Fractions with at least A 280 ϭ 0.05 were dialyzed for 24 h against 1ϫ PBS. The dialyzed fractions were concentrated using Microcon YM 30 (Millipore), and the protein concentration was determined using a Bio-Rad protein assay kit.
Isolation of Frog Oocytes-Oocytes were surgically removed from a female X. laevis obtained from Xenopus Express and tumbled with 2 mg/ml type I collagenase (Sigma) in Ca 2ϩ -free modified Barth's saline medium (20 mM HEPES, pH 7.5, 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 0.82 mM MgSO 4 ) at room temperature until most of the oocytes were free. The oocytes were washed thoroughly with this medium and stored overnight at 18°C in a modified Barth's saline medium (20 mM HEPES, pH 7.5, 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 0.82 mM MgSO 4, 0.33 mM Ca(NO 3 ) 2, 0.41 mM CaCl 2 ).
Microinjection of Antisense Oligodeoxynucleotides and Analysis of rRNA-Different antisense phosphodiester oligodeoxynucleotides (In-tegrated DNA Technologies, Inc.) with perfect complementation to both xGu-1 and xGu-2 mRNAs were used in this study (Table II). Stage V and VI oocytes were selected for cytoplasmic microinjection with 32.2 nl of antisense oligonucleotide (1 ng/nl). After 6 h of incubation at 18°C in modified Barth's saline medium, the oocytes were microinjected with a mixture of 16.1 nl of antisense oligonucleotide and 16.1 nl of [␣-32 P]GTP (3,000 Ci/mmol and 10 Ci/ul). After further incubation for 18 h at 18°C, the oocytes were homogenized in 50 l of homogenization buffer/ oocyte (50 mM NaCl, 50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 0.5% SDS, 200 g/ml proteinase K) and incubated at 37°C for 1 h. Samples were extracted with phenol:chloroform:iso-amyl alcohol (25:24:1) twice and with chloroform once. Nucleic acids were precipitated with ethanol. Total RNA equivalent to 1.5 oocytes was loaded per lane and resolved on a 1% agarose formaldehyde denaturing gel. Initial electrophoresis was done at 100 volts for 1 h and continued at 140 volts for 2 h. The gel was washed twice with water for 5 min followed with 10ϫ SSC for 15 min. RNA was passively transferred onto a nitrocellulose membrane with 10ϫ SSC overnight. The membrane was air-dried for 30 min and exposed to a phosphorus imaging screen (Molecular Dynamics). The results were obtained by scanning the screen using STORM 860 Phos-phorImager and analyzed with ImageQuant software (Molecular Dynamics).
The 5.8 S rRNA was analyzed on a 10% polyacrylamide/7 M urea gel. A 4-8 S HeLa RNA fraction prepared according to Reddy et al. (39) was used as nonradioactive markers. After electrophoresis, the denaturing gel was stained with 0.2% methylene blue for 10 min at room temperature, destained with water, vacuum dried, and analyzed by autoradiography.
Determination of Half-life-The oocytes were microinjected with 32.2 nl of antisense BV795 (1 ng/nl) and incubated in a modified Barth's saline medium containing 1 mg/ml cycloheximide to stop protein synthesis (40). In the event cycloheximide did not completely inhibit synthesis of xGu protein, the microinjection of BV795 would contribute to the inhibition of such synthesis by degradation of xGu mRNA. Ten nuclei were manually isolated after incubation of oocytes for 0, 2, 3, 4, and 5 h and used for immunoblot analysis. Signal intensities were quantitated using a densitometer and ImageQuant software.

FIG. 2. Analysis of xGu expression.
A, Northern blot analysis of total RNA isolated from frog oocytes using a 32 Plabeled xGu-1 probe. The numbers to the left are RNA molecular mass markers in kilobases. B, colorimetric Western blot analysis of oocyte extracts obtained using freon fractionation. An affinity-purified anti-xGu antibody (30 ng/l) was used. The numbers to the left are molecular mass markers in kilodaltons. C, RT-PCR analysis of xGu-1 and xGu-2 using total RNA isolated from different frog tissues. xGu-1 and xGu-2 were separately amplified using 100 ng total RNA as described under "Experimental Procedures." Different volumes of 5 and 20 l of RT-PCR products for xGu-1 and xGu-2, respectively, were analyzed on an agarose gel. The amplification of ␤-actin was used as an internal control.
FIG. 3. Cellular localization of xGu. A6 X. laevis kidney cells were stained by indirect immunofluorescence using antibodies against Xenopus nucleolin (C23) and Xenopus Gu (xGu) proteins. Antimouse IgG coupled to rhodamine and anti-rabbit IgG coupled to fluorescein isothiocyanate were used as secondary antibodies that recognize anti-C23 and anti-xGu, respectively. Hoechst 33258 (0.5 g/ml) was used to stain chromosomal DNA.
serts were sequenced to prove absence of mutations. The EcoRI-linearized constructs were used as templates to in vitro synthesize xGu-1 mRNA using the Ambion MEGAscript SP6 kit. The sizes of the RNA products were checked on a formaldehyde-agarose gel. The RNA products were then used as templates for in vitro translation, and the resulting proteins were analyzed by Western blot using anti-xGu antibody.
Stage V and VI oocytes were selected for cytoplasmic microinjection with 32.2 nl of 1 ng/nl BV795 antisense oligonucleotide. After 4 h at Microinjection of Anti-xGu Antibody-Stage V and VI oocytes were selected for nuclear microinjection with 18.4 nl of affinity-purified anti-xGu antibody (1 ng/nl). After 4 h at 18°C in modified Barth's saline medium, the oocytes were microinjected a second time with 9.2 nl of the same antibody together with 9.2 nl of [␣-32 P]GTP. The oocytes were further incubated at 18°C in modified Barth's saline medium for 18 -20 h prior to total RNA isolation and analysis.

Primary Structures of the Two X. laevis RNA Helicase II/Gu
Proteins-Two cDNA clones from X. laevis that are highly homologous to mammalian Gu␣ protein are identified in this study. The amino acid sequences deduced from these two cDNA clones are 90% identical, excluding the 44-amino acid region present only in xGu-2 (Fig. 1). The presence of two Gu genes in X. laevis with 10% sequence divergence is consistent with the pseudotetraploid genetic make-up of this species (43).
Comparison of the amino acid sequences of mouse Gu␣, Gu␤, and the two Gu sequences from X. laevis shows that xGu-1 and xGu-2 are both homologues of mouse Gu␣. A search of the current X. laevis EST Data Bank library did not reveal any Gu␤ sequence homologue. Fig. 1 shows the alignment of the cDNA-derived amino acid sequences of the Gu proteins from mouse and X. laevis. The eight motifs highly conserved among RNA helicases are almost identical between the mouse and frog homologues (boxed in Fig. 1). Both xGu-1 and xGu-2, like their mammalian homologues, have a DEVD (Asp-Glu-Val-Asp) motif that deviates from the prototype DEAD or DEAH sequence (44).
The N-terminal regions of the two xGu proteins have poor homology with the mouse homologue (Fig. 1). The putative initiation of translation in the two xGu proteins is a part of an MP(G/V)K sequence conserved in mammalian Gu␣ and Gu␤ (31). All known Gu sequences, including the ␣ and ␤ forms, have MPGK at the N terminus except for xGu-1, which has MPVK. The presence of an in-frame UAG stop codon 15 nucleotides upstream of the putative AUG initiation codon in xGu-1 suggests that this AUG is the actual initiation of translation. Further examination of the N-terminal sequence of the mouse Gu␣ protein indicates the presence of three 37-amino acid tandem repeats near the N-terminal end (amino acids 117-227) encoded by a single exon ( Fig. 1; see also Ref. 34). Although not homologous to the mouse Gu␣ repeats, xGu-2 contains two 38-amino acid tandem repeats near the N-terminal end (amino acids 81-118 and 125-162). These two repeats differ from each other by a single amino acid residue (Asn 116 versus Asp 160 ). xGu-1 has only one copy of this repeat (amino acids 82-118). The mouse Gu␣ repeats also contain a KSP(K/R)L motif, which is a putative phosphorylation site for a maturation-promoting factor (45). In contrast, the two xGu proteins do not contain a KSP(K/R)L motif. The biological function of this phosphorylation site remains to be identified as far as the mouse Gu␣ protein is concerned.
The C-terminal regions of mGu␣, xGu-1, and xGu-2 have higher homology than the N-terminal regions. The FRGQR repeats near the carboxyl end, previously shown to be critical for the RNA folding activity of the human Gu␣ (30), are present in the mouse homologue but not in the xGu enzymes. The carboxyl ends of the two xGu proteins, however, are rich in glycine and arginine residues, which may possess RNA binding activity.
Expression of xGu-1 and xGu-2-The Xenopus cDNA clones we identified are 2444 and 2607 nucleotides in length for xGu-1 and xGu-2, respectively. Northern blot analysis using RNA isolated from oocytes indicates the presence of an ϳ2.7-kb mRNA ( Fig. 2A). The presence of diffuse bands indicates heterogeneity in xGu mRNA.
Western blot analysis of oocyte extracts, using an affinitypurified antibody raised against recombinant xGu-1 that crossreacts with xGu-2, shows a 95-kDa protein (Fig. 2B). The calculated molecular masses of the two Xenopus Gu homologues are 89 and 85 kDa. The presence of a single protein band in the oocyte extracts (Fig. 2B) implies a negligible difference in the actual molecular masses of the two forms of xGu protein. However, we sometimes observed two closely migrating bands on Western blots depending on the source of the oocytes and preparation of the extracts, which we attributed to partial protein degradation (see below).
To determine the tissue-specific expression of xGu-1 and xGu-2, total RNAs were extracted from different tissues of X. laevis, and their levels of expression were analyzed by RT-PCR. Fig. 2C shows higher expression of xGu-1 in all tissues analyzed compared with xGu-2. After normalization of the results relative to ␤-actin, the highest expression of xGu-1 and xGu-2 was observed in the stomach. The lowest expressions of xGu-1 and xGu-2 were observed in the kidney and the lungs, respectively.
To determine the expression and cellular localization of the FIG. 5. Northern blot analysis to identify RNAs migrating ahead of 18 S. Total RNA from control oocytes or those microinjected with BV579 control oligonucleotide or BV795 antisense as described in the legend to Fig. 4 (but without [␣-32 P]GTP microinjection) was analyzed by probing with 32 P end-labeled 18 S-5Ј corresponding to the 5Ј end of 18 S rRNA (A). The same filter was stripped of the probe and hybridized sequentially with 28 S and ITS1 oligonucleotide probes (B and C). After the last hybridization, the positions of the 18 and 28 S rRNAs were determined by staining the nitrocellulose filters with methylene blue. two xGu proteins in A6 X. laevis kidney cells, affinity-purified anti-xGu antibody was used for indirect immunofluorescence. As a control, antibody against nucleolar protein C23 was used for comparison. Although Xenopus C23/nucleolin localizes throughout the nucleolus (Fig. 3, C23), xGu localizes to discrete subnucleolar regions that also contain C23 (Fig. 3, xGu and C23ϩxGu). The localization of xGu in nucleoli supports the notion that the X. laevis clones we isolated are the homologues of the mammalian Gu protein.
Oligodeoxynucleotides Antisense to xGu mRNA Affect rRNA Processing-The localization of xGu to nucleoli suggests a possible role in the production of rRNA. We used antisense phosphodiester oligonucleotides to down-regulate xGu expression and examine rRNA production. Thirteen 20-mer oligonucleotides with perfect complementary sequences to both xGu-1 and xGu-2 mRNAs (see "Experimental Procedures" for sequences) were divided into four groups as shown in Fig. 4A. Each group was microinjected into oocytes, and RNA was labeled with [␣-32 P]GTP. Because ϳ70% of cellular transcription is devoted to rRNA production, the majority of the labeled transcripts will be intermediate and mature products of rRNA biosynthesis. Total RNA was isolated and separated on a denaturing agarose gel. Microinjection of group C or D antisense did not affect the rRNA processing in frog oocytes (data not shown). Group A antisense had a minimal effect, but Group B significantly affected the processing of rRNA (Fig. 4B) compared with BV579 (5Ј-TTCAGGGACCGGCGAGATACC-3Ј), a negative control oligonucleotide corresponding to a promoter region of the mouse Gu␣ gene. Group B, which included BV795-BV798, caused an increase in the level of 20 S and an appearance of multiple RNA bands that migrated faster than 18 S (Fig. 4B, lanes 5 and 6). As a positive control to show the inhibition of oocyte rRNA processing, we used an antisense oligodeoxynucleotide (5Ј-GCTGTT-TCTC-3Ј; see Fig. 4, B, lane 7 and C, lane 16) that targeted Xenopus U8 small nucleolar RNA and was previously shown to inhibit the processing of a 36 S rRNA intermediate (36).
To determine which specific oligonucleotides in Group B affected rRNA processing, BV795-BV798 were individually microinjected into frog oocytes. Fig. 4C shows that microinjection of BV795 and BV796 (lanes 12-15) but not BV797 and BV798 FIG. 6. Analysis of xGu expression. A, RNA samples that were prepared as described in the legend to Fig. 4 were analyzed by RT-PCR using primers that amplify both Xenopus Gu homologues (xGu), nucleolin (xC23), nucleophosmin (xB23), and ␤-actin. RT-PCR products on agarose gels were stained with ethidium bromide. Similar samples were analyzed by RNase protection assay. B, to determine the kinetics of xGu mRNA degradation, total RNA was isolated from oocytes at different time points after a one-time microinjection with BV795 antisense and analyzed for the level of expression of xGu and xB23. C, the level of xGu protein was determined by isolating germinal vesicles from nonmicroinjected oocytes (No Oligo) or 24 h after microinjection with BV579 or BV795 antisense and then boiling in Laemmli buffer. Each lane was loaded with protein extracts equivalent to 3.75 oocytes and analyzed by Western blot using antibodies against xGu and nucleolin (xC23). Another gel that was loaded with equivalent amounts of proteins was stained with Coomassie Blue. D, the half-life of xGu was determined as described under "Experimental Procedures." The relative intensities are plotted on a semilogarithmic graph.
(lanes 10 and 11) resulted in a similar pattern of aberrant rRNA processing compared with that found with Group B. The results suggest that BV795 and BV796 are more efficient than other oligonucleotides at targeting xGu, resulting in inhibition of the production of mature 18 and 28 S rRNAs.
The decrease in the level of 18 S, with a concomitant increase in 20 S rRNA, suggests inhibition of the processing of 20 to 18 S rRNA. The decrease in the amount of 28 S rRNA, however, may not be due to a defective processing of its direct precursor, 32 S rRNA, as shown by the absence of 32 S accumulation (Fig.  4C, lanes 12-15), and this suggests a possible degradation of 28 S, which could account for the appearance of RNA fragments migrating faster than 18 S. However, the results may also suggest degradation of 28 S precursors. To determine whether 36 S is degraded in the presence of xGu antisense oligonucleotide, the oocytes were microinjected with U8 antisense resulting in the accumulation of 36 S with almost complete inhibition of 28 S formation but unaffected 18 S (Fig. 4D, lane 4). When the U8 antisense-treated oocyte was microinjected with BV795, the levels of 40S and 36 S were not affected, and smaller fragments below 18 S were not observed (Fig. 4D, lane 5). The results suggest that 40 and 36 S rRNAs are not degraded in the presence of BV795 and that the degradation products seen below 18 S must have come from other rRNAs. Consistent with the data in Fig. 4C, microinjection of BV795 in U8 antisensetreated oocytes caused an increase in the level of 20 S and a decrease in 18 S rRNA (compare lanes 4 and 5 in Fig. 4D).
To determine whether 32 S rRNA is degraded in the presence of BV795 or BV796, we analyzed the changes in the level of 5.8 S. Our hypothesis was that degradation of 32 S would result in a lower amount of 5.8 S. Fig. 4E shows that microinjection of BV795 and BV796 (lanes 5-8) did not affect the level of 5.8 S, suggesting normal processing of 32 S.
Overall, the results indicated that microinjection of BV795 and BV796 antisense oligonucleotides resulted in the degradation of 28 S rRNA. To further prove this hypothesis, an RNA blot analysis was done using 32 P-labeled oligonucleotides corresponding to different regions of 18 and 28 S as hybridization probes. Fig. 5A shows that the 18 S 5Ј probe hybridized to 18 S rRNA but did not hybridize with any faster migrating fragments. A similar result was obtained using an 18 S 3Ј probe (data not shown). A 28 S 5Ј probe hybridized with the 28 S rRNA and the fragments migrating faster than 18 S (Fig. 5B). These fragments are products of the normal degradation of 28 S rRNA, and the process is enhanced in the presence of BV795 (compare control lanes NM and BV579 with BV795 in Fig. 5B).
The results indicate that the RNA fragments below 18 S on the gel are mostly degradation products of 28 S rRNA.
It is not known why 18 S 5Ј probe did not hybridize with 20 S rRNA under our hybridization conditions (Fig. 5A). However, a mixture of ITS1A and ITS1B probes (antisense to internal transcribed sequence 1) identified 20 S rRNA (Fig. 5C). These probes show an increase in the level of 20 S rRNA in oocytes microinjected with BV795 antisense oligonucleotide, whereas 18 and 28 S probes show a decrease in 18 and 28 S rRNA, respectively. It should be noted that all of these probes identify both newly synthesized and old rRNAs and show minimal changes in the level of rRNAs compared with the larger differences observed in the level of 32 P-labeled newly synthesized rRNAs (Fig. 4).
Antisense-mediated Decrease in xGu mRNA and Protein Levels-To determine whether antisense BV795 specifically inhibits expression of xGu, RT-PCR and RNase protection assays were carried out to measure changes in the mRNA levels. Primers for RT-PCR were designed to amplify both xGu-1 and xGu-2 as indicated by the presence of two DNA bands (Fig. 6A). RT-PCR showed an 89% decrease in the total xGu mRNA level when oocytes were microinjected with BV795 compared with oocytes microinjected with a control oligonucleotide BV579, which is not related to xGu (Fig. 6A). xGu-specific BV795 antisense did not affect the level of frog nucleolin (C23) or nucleophosmin (B23) mRNA, two other nucleolar proteins known to play important roles in rRNA processing (46 -48). The BV795-mediated decrease in xGu mRNA was further supported by an RNase protection assay. We designed a probe to determine changes in the steady-state level of xGu-1 mRNA using the RNase protection assay. A 90% decrease in the intensity of the protected band after microinjection with BV795 was observed consistent with RT-PCR results.
The degradation of xGu mRNA after antisense BV795 microinjection was also examined. One hour after microinjection, a 47% reduction in total xGu mRNA was observed (Fig. 6B). The rate of degradation, however, decreased with time, which might be due to degradation of BV795 oligonucleotide. It should be noted that RT-PCR showed a total decrease in xGu-1 and xGu-2 mRNA. We also examined the protein level for xGu after microinjection of BV795. Western blot analysis of nuclear extracts from oocytes microinjected with BV795 showed a 90% decrease in xGu protein relative to oocytes microinjected with control oli- FIG. 7. Reversal of the effects mediated by BV795 antisense. Total RNA isolated from oocytes microinjected with the indicated oligonucleotide and [␣-32 P]GTP was analyzed by gel electrophoresis (A) as described in the legend to Fig. 4 or by RT-PCR (B). C, protein extracts were analyzed using a chemiluminescence detection system and antibodies against xGu and nucleolin (xC23). In all panels, lane 1 refers to an absence of the first initial microinjection and then a later microinjection with water; lane 2 shows microinjection with BV579 negative control oligonucleotide, followed by microinjection with water; lane 3 shows microinjection with BV795 antisense, followed by microinjection with water; and lane 4 shows microinjection with BV795 antisense, followed by microinjection with xGu-1 mRNA. gonucleotide BV579 (Fig. 6C). The results indicate that microinjection of oocytes with BV795 antisense results in a specific decrease in the mRNA and protein levels of xGu without a significant effect on the levels of other nucleolar proteins. The rapid degradation of the protein occurs within 24 h after the first microinjection. Fig. 6D shows a half-life of 2.5 h, which implies that most xGu protein that was in the oocyte before antisense microinjection would have been degraded prior to the extraction of total RNA. The short half-life of xGu is somewhat surprising, but similar short half-lives of 3-4 h have been reported for the subunits of the epithelial sodium channels in Xenopus oocytes (49) and 40 -50 min in A6 frog kidney cell line (50).
Microinjection of in Vitro Synthesized xGu-1 mRNA Reversed the Antisense Effects-To determine whether the BV795 antisense-mediated aberrance in rRNA processing could be rescued, an in vitro synthesized xGu-1 mRNA was microinjected into the oocytes previously treated with BV795. Fig. 7A shows antisense-mediated defects in rRNA processing (lane 3), which are reversed when the antisense-treated oocytes were microinjected with in vitro synthesized xGu-1 mRNA (lane 4).
To analyze the mRNA levels of xGu-1 and xGu-2, RT-PCR was carried out using a pair of primers that amplified both homologues. Fig. 7B shows that BV795 antisense decreased both xGu-1 and xGu-2 mRNAs (lane 3) without any effects on nucleophosmin (xB23). Microinjection of the in vitro synthesized xGu-1 mRNA increased the level of xGu-1 mRNA but not xGu-2, as expected (lane 4). Western blot analysis shows an increase in xGu protein upon microinjection of in vitro synthesized xGu-1 mRNA into BV795-treated oocytes (Fig. 7C, lane  4). Overall, the results indicate that a decrease in the expression level of xGu results in defective rRNA processing, which can then be reversed by introducing an exogenous xGu mRNA.
The xGu DEVD Mutant Lacks RNA Unwinding Activity, but Introduces Secondary Structure to ssRNA, and Is Unable to Reverse the Antisense Effects-The human Gu␣ homologue possesses ATPase, RNA unwinding, and RNA folding activities in vitro (28 -30). It requires Mg 2ϩ and ATP for its RNA helicase but not for its foldase activity. Except for dATP, other nucleotides cannot substitute the required ATP for its RNA unwinding activity. GTP slightly inhibits its helicase function but stimulates the RNA folding activity of the human homologue (51). Enzyme assays of the wild type xGu protein fused to GST showed an RNA helicase activity comparable with its human homologue (Fig. 8). The enzymes human Gu␣ and Xenopus Gu could unwind a double-stranded RNA with 5Ј overhang with FIG. 8. Characterization of the wild type and mutant forms of xGu. A, purified bacterially expressed xGu wild type and mutant forms (0.6 g each) were analyzed on a 10% polyacrylamide-SDS gel and stained with Coomassie blue. B, for the helicase assay a double-stranded RNA (dsRNA) with 5Ј overhang and 32 P-labeled lower strand was obtained by annealing two transcripts synthesized using BamHI-and HindIII-digested pBluescript II KS plasmid. For the RNA folding assay the same 32 P-labeled lower strand was used as ssRNA substrate. These substrates were used to assay the RNA unwinding and RNA folding activities of the recombinant GST fusion proteins (50 ng each) of the human (hGu) and Xenopus (xGu) Gu homologues. The mutant form of xGu was assayed using 50 and 100 ng of fusion protein. C, the ability of the xGu mutant form to reverse the BV795-mediated inhibitory effects on rRNA production was determined as described in the legend to Fig. 7. The bottom panels show the levels of xGu and B23 mRNA determined by RT-PCR. M, molecular mass markers in kilodaltons; WT, wild type; Mut, DEVD mutant; dpRNA, displaced RNA; fRNA, folded RNA. almost identical efficiencies. Mutation of the DEVD motif of xGu to ASVD completely inhibited its RNA unwinding activity, and a more slowly gel-migrating RNA band was produced by the assay (Fig. 8B). The more slowly migrating RNA band cannot be due to xGu-RNA complexes because the reaction mixtures were treated with proteinase K immediately after the helicase assay, and the samples were resolved in polyacrylamide gels containing SDS. The identity and structure of this more slowly migrating RNA band remain to be identified.
The cDNA-derived amino acid sequence of xGu does not have the FRGQR repeats seen in its human orthologue (Fig. 1). This domain is responsible for the ability of human Gu␣ to introduce secondary structure to ssRNA (30). Fig. 8B shows that the wild type GST-xGu fusion protein is less active in folding an ssRNA as compared with its human homologue. However, what is interesting is that inactivation of the helicase activity of GST-xGu (DEVD mutant) resulted in an enhanced ability to introduce secondary structure. A similar observation was obtained when the RNA folding activities of human Gu␣ and its DEVD mutant were compared (51). It remains to be determined whether the arginine-glycine-rich C terminus of xGu is responsible for the RNA folding activity.
Microinjection of the xGu mutant into BV795-treated oocytes did not rescue the aberrant rRNA processing, whereas the wild type form did reverse this antisense effect, consistent with the results in Fig. 7 (Fig. 8C). The results imply that the helicase activity of xGu is important for the production of 28 and 18 S rRNAs.
Anti-xGu Antibody Inhibits 18 and 28 S rRNA Formation-We previously reported that binding of anti-Gu␣ antibody to a recombinant human Gu␣ resulted in the inhibition of its RNA helicase activity in vitro (51). We hypothesized that microinjection of affinity-purified anti-xGu antibody into X. laevis oocytes would have a similar inhibitory effect on the RNA unwinding activity of xGu and consequently inhibit rRNA processing. Fig. 9 shows that microinjection of affinity-purified anti-xGu antibody results in decrease of 18 and 28 S rRNAs and the accumulation of 20 S and smaller RNA fragments migrating ahead of 18 S, all of which is similar to the observed effects of BV795 antisense on rRNA processing (Fig. 4). In contrast, microinjection of rabbit IgG into oocytes did not result in aberrant rRNA processing. The results suggest that antibody neutralization of xGu and antisense-mediated down-regulation of xGu expression both result in the same inhibitory effects on the production of 18 S rRNA and induce instability in 28 S rRNA.
Localization of Gu␣ in the DFC and GC-Early and late processing of rRNA intermediates occurs in the DFC and GC of the nucleolus, respectively. If Gu␣ protein is involved in these processes, we hypothesized that it would be localized to the DFC and GC, but not to the FC. Immunoelectron microscopy using rat kangaroo kidney epithelial cells (PtK2) shows a preferential localization of Gu␣ protein to the DFC. Gu␣ protein also localizes to the GC to a lesser extent but not to the FC (Fig.  10). Although this observation is made in a rat kangaroo cell line, the result is consistent with the identified function of xGu in the production of mature 18 and 28 S rRNAs considering the highly conserved rRNA processing through evolution. DISCUSSION The present study demonstrates a function for the Gu␣ Xenopus amphibian homologues in the production of rRNA. The two Xenopus homologues xGu-1 and xGu-2, which are identified in this study, produce nearly identical proteins and are referred to as one protein in this discussion. The two gene products may have identical functions that are based on three major observations: 1) functional analysis using antisense oligonucleotide to down-regulate xGu expression shows degradation of 28 S to smaller RNA fragments and inhibition of the processing of 20 to 18 S rRNA, 2) microinjection of anti-xGu antibody, which inhibits xGu enzymatic activity in vitro, shows similar inhibitory effects on rRNA processing as that observed in the antisense experiments, and 3) immunoelectron microscopy shows Gu protein localized to the dense fibrillar and granular components of the nucleolus, sites of early and late processing of rRNA.
Functional analysis of xGu enzyme in X. laevis oocytes was carried out by down-regulating its gene expression. Two antisense oligonucleotides complementary to both xGu-1 and xGu-2 specifically degraded xGu mRNA (Fig. 6), presumably mediated by RNase H (52). The degradation of xGu mRNA and a concomitant decrease in xGu protein occurred quickly and may have preceded the observed aberrant processing of rRNA. Based on these results, xGu seems to function at two important steps in the biogenesis of rRNA as shown in Fig. 11: processing of 20 to 18 S rRNA as well as playing a role in the stabilization of 28 S rRNA. In general, the processing of 20 to 18 S rRNA involves endonucleolytic cleavages to release the external transcribed spacer 1 (ETS1) and internal transcribed spacer 1 (ITS1), leading to a mature 18 S rRNA in the nucleolus (53). Little is known about such processing events in higher eukaryotes. In yeast, the 20 S pre-rRNA contains 18 S rRNA and the 5Ј region of the ITS1. It is believed that the yeast 20 S rRNA is exported to the cytoplasm where it is processed into mature 18 S rRNA (54,55). There is no evidence that a similar processing of 20 S pre-rRNA in higher eukaryotes occurs in the cytoplasm. Our results support the idea that the metazoan 20 S rRNA is processed in the nucleolus and that xGu is involved.
What then could be the specific function of xGu in the processing of 20 S rRNA? Gu might be a component of a complex ribonucleoprotein (RNP) particle and may mediate structural organization of RNA-protein and RNA-RNA complexes in these RNP particles (27,56,57). This hypothesis is supported by the fact that human Gu␣ is found in large RNP particles separated on sucrose density gradients, and treatment of the nuclear extracts with RNase A destroys the integrity of the particles and releases Gu␣ protein. 2 Furthermore, three recent independent proteomic studies (58 -60) show that Gu␣ is a component of the preribosomal processing machinery. In the previous studies, different FLAG-tagged nucleolar proteins were used to pull down protein complexes, and the protein components were 2 D. Henning and B. C. Valdez, unpublished results.
FIG. 9. Effects of anti-xGu antibody on rRNA processing. Total RNA isolated from control oocytes and oocytes microinjected with rabbit IgG or affinity-purified anti-xGu polyclonal antibody was analyzed as described in the legend to Fig. 4. NM, no microinjection.
identified by mass spectrometry. Among the identified proteins were Gu␣ and other nucleolar proteins like nucleolin, suggesting that Gu␣ is a component of a common RNP involved in preribosomal RNA processing.
It is still not known how rRNA is organized into a defined complex in the nucleolus, but small nucleolar RNAs and nonribosomal proteins are likely to be involved in pre-rRNA processing by aiding in proper folding of the RNA components. For instance, 18 and 28 S rRNAs that have the incorrect conformation probably do not associate with the correct ribosomal proteins and thus become substrates for degradation by the exosome (61). The RNA helicase activity of xGu might aid in maintaining a proper conformation of rRNA so that proper assembly with ribosomal proteins can occur (62,63). Our data point to the possible involvement of xGu in the maintenance of the proper conformation of 28 S rRNA. For example, depletion of xGu protein leads to degradation of 28 S rRNA. An unpublished observation from our laboratory shows an interaction between human Gu␣ and the ribosomal protein L4 (RPL4). The bacterial homologue of RPL4 is known to bind to 23 S rRNA (64), which is equivalent to the mammalian and metazoan 28 S rRNA. Perhaps Gu␣ protein rearranges 28 S rRNA conformation prior to binding of RPL4 to 28 S rRNA within an RNP particle. Consequently, the absence of Gu␣ may not allow RPL4 and/or other factors to bind to 28 S rRNA.
In this paper we have shown that the depletion of 18 and 28 S rRNA mediated by xGu antisense in frog oocytes is reversed in the presence of in vitro synthesized wild type xGu mRNA but not its mutant form. Mutation of the DEVD motif completely inhibits the in vitro RNA helicase activity of xGu (Fig. 8B). This mutant does not reverse the inhibition of rRNA production caused by BV795 antisense underlying the relevance of xGu RNA helicase activity (Fig. 8C). Although the mutant form possesses higher RNA folding activity than wild type (Fig. 8B), it remains to be determined whether abrogation of the RNA folding activity has any effect on rRNA production. This model system provides a means to dissect the structure-function relationship in xGu in relation to rRNA production.
The specificity of the xGu antisense-mediated effect is further supported by the microinjection of affinity-purified polyclonal anti-xGu antibody (Fig. 9). The binding of this antibody to xGu could have inhibited the enzymatic activity of this protein because we previously reported inhibition of the human Gu␣ RNA helicase activity by anti-Gu␣ antibody (51). Overall, down-regulation of xGu expression using antisense oligonucleotides or antibody neutralization resulted in the depletion of 18 and 28 S rRNAs in frog oocytes. Such involvement of Gu in the processing of pre-rRNA is consistent with its localization to the DFC and GC regions of the nucleolus (Fig. 10).
This study is the first report demonstrating direct involvement of a DEAD box protein in rRNA processing in a metazoan system. Moreover, we have identified specific sites where xGu functions in the production of mature rRNA. Bop1 is another mammalian protein (that is not a member of the DEAD box family) shown to coordinate processing of the spacer regions in mouse pre-rRNA (65,66). Other mammalian nucleolar proteins implicated in rRNA processing include p120 (67), nucleolin/C23 (47,48,68), nucleophosmin/B23 (46), and fibrillarin (53,69). Genetic and biochemical studies have previously identified the specific sites where the yeast homologues of these proteins function. For example, Nop2p, a p120 yeast homologue, is involved in the processing of 27S pre-rRNA to the mature 25 S rRNA (70). Disruption of nsr1 gene, the yeast homologue of C23, results in underaccumulation of the 18 S rRNA and reduction in A o , A 1 , and A 2 cleavages (19,71).
The yeast homologue of xGu remains to be identified. Comparison of the xGu-1 amino acid sequence against all known yeast nucleolar RNA helicases shows Rrp3p to have 31% (137/ 435) identity and 45% (202/435) similarity to xGu in the region where RNA helicase motifs are conserved. Depletion of Rrp3p in yeast results in inhibition of 18 S rRNA production and inhibition of endonucleolytic cleavages at the A 1 and A 2 sites (72). It remains to be determined whether xGu can complement depletion of the Rrp3p protein in yeast to prove whether Rrp3p is the xGu orthologue.
Our present and previous findings indicate that Gu is a multifunctional RNA helicase protein. Its interaction with c-Jun may up-regulate expression of proteins relevant to the processing and stability of rRNAs. Such a function may augment a more direct role for Gu in rRNA biosynthesis, as reported in the present study. Although a possibility exists that the observed antisense-mediated down-regulation of xGu could have resulted in low expression of other nucleolar proteins involved in rRNA production, our present data support a direct role for xGu in rRNA production, as shown by its localization in specific nucleolar sites known to be involved in rRNA synthesis and processing, its presence in preribosomal RNPs, and its interaction with ribosomal protein L4.
In summary, evidence is presented that shows the involvement of xGu in the processing of 20 to 18 S rRNA, as well as a role in the stabilization of the 28 S rRNA. Our focus now is to determine whether or not the mammalian homologue has similar physiological functions. Our rescue experiment, in which we introduced the in vitro synthesized xGu mRNA, provides the groundwork to identify which domains of xGu are involved in specific steps of rRNA biosynthesis.