Nucleotide sequence determining the first cleavage site in the processing of mouse precursor rRNA

Che-1/AATF binds to RNA polymerase I machinery and sustains ribosomal RNA gene transcription

Originally identified as an RNA polymerase II interactor, Che-1/AATF (Che-1) has now been recognized as a multifunctional protein involved in cell-cycle regulation and cancer progression, as well as apoptosis inhibition and response to stress. This protein displays a peculiar nucleolar localization and it has recently been implicated in pre-rRNA processing and ribosome biogenesis. Here, we report the identification of a novel function of Che-1 in the regulation of ribosomal RNA (rRNA) synthesis, in both cancer and normal cells. We demonstrate that Che-1 interacts with RNA polymerase I and nucleolar upstream binding factor (UBF) and promotes RNA polymerase I-dependent transcription. Furthermore, this protein binds to the rRNA gene (rDNA) promoter and modulates its epigenetic state by contrasting the recruitment of HDAC1. Che-1 downregulation affects RNA polymerase I and UBF recruitment on rDNA and leads to reducing rDNA promoter activity and 47S pre-rRNA production. Interestingly, Che-1 depletion induces abnormal nucleolar morphology associated with re-distribution of nucleolar proteins. Finally, we show that upon DNA damage Che-1 re-localizes from rDNA to TP53 gene promoter to induce cell-cycle arrest. This previously uncharacterized function of Che-1 confirms the important role of this protein in the regulation of ribosome biogenesis, cellular proliferation and response to stress.

PTRF/Cavin-1 promotes efficient ribosomal RNA transcription in response to metabolic challenges

Ribosomal RNA transcription mediated by RNA polymerase I represents the rate-limiting step in ribosome biogenesis. In eukaryotic cells, nutrients and growth factors regulate ribosomal RNA transcription through various key factors coupled to cell growth. We show here in mature adipocytes, ribosomal transcription can be acutely regulated in response to metabolic challenges. This acute response is mediated by PTRF (polymerase I transcription and release factor, also known as cavin-1), which has previously been shown to play a critical role in caveolae formation. The caveolae–independent rDNA transcriptional role of PTRF not only explains the lipodystrophy phenotype observed in PTRF deficient mice and humans, but also highlights its crucial physiological role in maintaining adipocyte allostasis. Multiple post-translational modifications of PTRF provide mechanistic bases for its regulation. The role of PTRF in ribosomal transcriptional efficiency is likely relevant to many additional physiological situations of cell growth and organismal metabolism.

DOI:http://dx.doi.org/10.7554/eLife.17508.001

Obesity can cause several other health conditions to develop. Type 2 diabetes is one such condition, which arises in part because fat cells become unable to store excess fats. This makes certain tissues in the body less sensitive to the hormone insulin, and so the individual is less able to adapt to changing nutrient levels. Without treatment or a change in lifestyle, this insulin resistance may develop into diabetes. However, “healthy obese” individuals also exist, who can accommodate an overabundance of fat without developing insulin resistance and diabetes.

Some forms of rare genetic disorders called lipodystrophies, which result in an almost complete lack of body fat, can also lead to type 2 diabetes. This raises the question of whether lipodystrophy and obesity share some common mechanisms that cause fat cells to trigger insulin resistance. One possible player in such mechanisms is a protein called PTRF. In rare cases, individuals with lipodystrophy lack this protein, and mice that have been engineered to lack PTRF also largely lack body fat and develop insulin resistance.

Fat cells can respond rapidly to changes in nutrients during feeding or fasting, and to do so, they must produce new proteins. Structures called ribosomes, which are made up of proteins and ribosomal RNA, build proteins; thus when the cell needs to make new proteins, it also has to produce more ribosomes. PTRF is thought to play a role in ribosome production, but it is not clear how it does so.

Liu and Pilch analyzed normal mice as well as those that lacked the PTRF protein. This revealed that in response to cycles of fasting and feeding, PTRF increases the production of ribosomal RNA in fat cells, enabling the cells to produce more proteins. By contrast, the fat cells of mice that lack PTRF have much lower levels of ribosomal RNA and proteins.

Liu and Pilch then examined mouse fat cells that were grown in the laboratory. Exposing these cells to insulin caused phosphate groups to be attached to the PTRF proteins inside the cells. This modification caused PTRF to move into the cell’s nucleus, where it increased the production of ribosomal RNA.

Overall, the results show that fat cells that lack PTRF are unable to produce the proteins that they need to deal with changing nutrient levels, leading to an increased likelihood of diabetes. The next steps are to investigate the mechanism by which PTRF is modified, and to see whether the mechanisms uncovered in this study also apply to humans.

DOI:http://dx.doi.org/10.7554/eLife.17508.002

Certain adenylated non-coding RNAs, including 5' leader sequences of primary microRNA transcripts, accumulate in mouse cells following depletion of the RNA helicase MTR4

RNA surveillance plays an important role in posttranscriptional regulation. Seminal work in this field has largely focused on yeast as a model system, whereas exploration of RNA surveillance in mammals is only recently begun. The increased transcriptional complexity of mammalian systems provides a wider array of targets for RNA surveillance, and, while many questions remain unanswered, emerging data suggest the nuclear RNA surveillance machinery exhibits increased complexity as well. We have used a small interfering RNA in mouse N2A cells to target the homolog of a yeast protein that functions in RNA surveillance (Mtr4p). We used high-throughput sequencing of polyadenylated RNAs (PA-seq) to quantify the effects of the mMtr4 knockdown (KD) on RNA surveillance. We demonstrate that overall abundance of polyadenylated protein coding mRNAs is not affected, but several targets of RNA surveillance predicted from work in yeast accumulate as adenylated RNAs in the mMtr4KD. microRNAs are an added layer of transcriptional complexity not found in yeast. After Drosha cleavage separates the pre-miRNA from the microRNA's primary transcript, the byproducts of that transcript are generally thought to be degraded. We have identified the 5′ leading segments of pri-miRNAs as novel targets of mMtr4 dependent RNA surveillance.

Construction of synthetic nucleoli in human cells reveals how a major functional nuclear domain is formed and propagated through cell division

Human cell nuclei are functionally organized into structurally stable yet dynamic bodies whose cell cycle inheritance is poorly understood. Here, we investigate the biogenesis and propagation of nucleoli, sites of ribosome biogenesis and key regulators of cellular growth. Nucleolar and cell cycles are intimately connected. Nucleoli disappear during mitosis, reforming around prominent uncharacterized chromosomal features, nucleolar organizer regions (NORs). By examining the effects of UBF depletion on both endogenous NORs and synthetic pseudo-NORs, we reveal its essential role in maintaining competency and establishing a bookmark on mitotic NORs. Furthermore, we demonstrate that neo-NORs, UBF-binding site arrays coupled with rDNA transcription units, direct the de novo biogenesis of functional compartmentalized neonucleoli irrespective of their site of chromosomal integration. For the first time, we establish the sequence requirements for nucleolar biogenesis and provide proof that this is a staged process where UBF-dependent mitotic bookmarking precedes function-dependent nucleolar assembly.

Functional dichotomy of ribosomal proteins during the synthesis of mammalian 40S ribosomal subunits

Subsets of 40S ribosomal subunits are required for initiating rRNA processing, rRNA maturation, and nuclear export.

Our knowledge of the functions of metazoan ribosomal proteins in ribosome synthesis remains fragmentary. Using siRNAs, we show that knockdown of 31 of the 32 ribosomal proteins of the human 40S subunit (ribosomal protein of the small subunit [RPS]) strongly affects pre–ribosomal RNA (rRNA) processing, which often correlates with nucleolar chromatin disorganization. 16 RPSs are strictly required for initiating processing of the sequences flanking the 18S rRNA in the pre-rRNA except at the metazoan-specific early cleavage site. The remaining 16 proteins are necessary for progression of the nuclear and cytoplasmic maturation steps and for nuclear export. Distribution of these two subsets of RPSs in the 40S subunit structure argues for a tight dependence of pre-rRNA processing initiation on the folding of both the body and the head of the forming subunit. Interestingly, the functional dichotomy of RPS proteins reported in this study is correlated with the mutation frequency of RPS genes in Diamond-Blackfan anemia.

A protein inventory of human ribosome biogenesis reveals an essential function of exportin 5 in 60S subunit export

The assembly of ribosomal subunits in eukaryotes is a complex, multistep process so far mostly studied in yeast. In S. cerevisiae, more than 200 factors including ribosomal proteins and trans-acting factors are required for the ordered assembly of 40S and 60S ribosomal subunits. To date, only few human homologs of these yeast ribosome synthesis factors have been characterized. Here, we used a systematic RNA interference (RNAi) approach to analyze the contribution of 464 candidate factors to ribosomal subunit biogenesis in human cells. The screen was based on visual readouts, using inducible, fluorescent ribosomal proteins as reporters. By performing computer-based image analysis utilizing supervised machine-learning techniques, we obtained evidence for a functional link of 153 human proteins to ribosome synthesis. Our data show that core features of ribosome assembly are conserved from yeast to human, but differences exist for instance with respect to 60S subunit export. Unexpectedly, our RNAi screen uncovered a requirement for the export receptor Exportin 5 (Exp5) in nuclear export of 60S subunits in human cells. We show that Exp5, like the known 60S exportin Crm1, binds to pre-60S particles in a RanGTP-dependent manner. Interference with either Exp5 or Crm1 function blocks 60S export in both human cells and frog oocytes, whereas 40S export is compromised only upon inhibition of Crm1. Thus, 60S subunit export is dependent on at least two RanGTP-binding exportins in vertebrate cells.

Characterization of the angiogenic activity of zebrafish ribonucleases

Ribonucleases identified from zebrafish possess angiogenic and bactericidal activities. Zebrafish RNases have three intramolecular disulfide bonds, a characteristic structural feature of angiogenin, different from the typical four disulfide bonds of the other members of the RNase A superfamily. They also have a higher degree of sequence homology to angiogenin than to RNase A. It has been proposed that all RNases evolved from these angiogenin-like progenitors. In the present study, we characterize, in detail, the function of zebrafish RNases in various steps in the process of angiogenesis. We report that zebrafish RNase-1, -2 and -3 bind to the cell surface specifically and are able to compete with human angiogenin. Similar to human angiogenin, all three zebrafish RNases are able to induce phosphorylation of extracellular signal-regulated kinase 1/2 mitogen-activated protein kinase. They also undergo nuclear translocation, accumulate in the nucleolus and stimulate rRNA transcription. However, zebrafish RNase-3 is defective in cleaving rRNA precursor, even though it has been reported to have an open active site and has higher enzymatic activity toward more classic RNase substrates such as yeast tRNA and synthetic oligonucleotides. Taken together with the findings that zebrafish RNase-3 is less angiogenic than zebrafish RNase-1 and -2 as well as human angiogenin, these results suggest that zebrafish RNase-1 is the ortholog of human angiogenin and that the ribonucleolytic activity of zebrafish RNases toward the rRNA precursor substrate is functionally important for their angiogenic activity.

Direct regulation of rRNA transcription by fibroblast growth factor 2

Fibroblast growth factor 2 (FGF-2), which is highly expressed in developing tissues and malignant cells, regulates cell growth, differentiation, and migration. Five isoforms (18 to approximately 34 kDa) of FGF-2 are derived from alternative initiation codons of a single mRNA. The 18-kDa FGF-2 isoform is released from cells by a nonclassical secretory pathway and regulates gene expression by binding to cell surface receptors. This isoform also localizes to the nucleolus, raising the possibility that it may directly regulate ribosome biogenesis, a rate-limiting process in cell growth. Although several growth factors have been shown to accumulate in the nucleolus, their function and mechanism of action remain unclear. Here we show that 18-kDa FGF-2 interacts with upstream binding factor (UBF), an architectural transcription factor essential for rRNA transcription. The maximal activation of rRNA transcription in vitro by 18-kDa FGF-2 requires UBF. The 18-kDa FGF-2 localizes to rRNA genes and is necessary for the full activation of pre-rRNA synthesis in vivo. Our results demonstrate that 18-kDa FGF-2 directly regulates rRNA transcription.

Estimation of ribosomal RNA transcription rate in situ

Traditionally, the rate of transcription is measured by metabolic labeling (e.g., the run-on assay), which can be carried out only in isolated or cultured cells. It has been difficult if not impossible to assess the rate of transcription of a gene in a specific cell type in situ. We show here that the quantity of 47S precursor ribosomal RNA (pre-rRNA), which correlates positively with the rate of rRNA transcription as measured by the run-on assay, can serve as an indicator for the rate of its transcription. We adopted this method as an in situ hybridization procedure to demonstrate its validity in vivo. The notion of using the quantity of the primary transcript as an indicator of the rate of transcription has the potential application in monitoring the rate of messenger RNA transcription in single cells within a tissue of complex cellular composition.

Phosphatidylinositol 3-kinase and mTOR signaling pathways regulate RNA polymerase I transcription in response to IGF-1 and nutrients

Regulation of ribosomal RNA gene transcription by RNA polymerase I (Pol I) is fundamental to ribosome biogenesis and therefore protein translation capacity and cell growth, yet little is known of the key signaling cascades involved. We show here that insulin-like growth factor-1 (IGF-1)-induced Pol I transcription in HEK293 cells is entirely dependent on phosphatidylinositol 3-kinase (PI3K) activity and, additionally, is modulated by the mammalian target of rapamycin (mTOR), which coordinates Pol I transcription with the availability of amino acids. The mitogen-activated protein kinase (MAPK) pathway is weakly stimulated by IGF-1 in these cells and partly contributes to Pol I transcription regulation. Activation of Pol I transcription by IGF-1 results from enhancement of the activity of the Pol I transcription machinery and increased occupancy by SL1 of the endogenous tandemly repeated ribosomal promoters in vivo. The inputs from PI3K, mTOR, and MAPK pathways converge to direct appropriate rRNA gene expression by Pol I in the nucleolus of mammalian cells in response to environmental cues, such as growth factors and nutrients.

Trypanosoma brucei 5'ETS A'-cleavage is directed by 3'-adjacent sequences, but not two U3 snoRNA-binding elements, which are all required for subsequent pre-small subunit rRNA processing events

Trypanosoma brucei pre-rRNA processing commences by cleavage near the 5' end of 5.8 S sequences. The 5' external transcribed spacer (5'ETS) is removed from pre-small subunit (SSU) rRNAs by sequential cleavages at internal A' and A0 sites, and A1 at the 5' end of SSU rRNA. The A' and A0 sites positionally resemble the U3 small nucleolar RNA-dependent, primary pre-rRNA cleavages of vertebrates and yeast, respectively. Uniquely in T. brucei, two U3-crosslinkable 5'ETS sites are essential for SSU rRNA production: site1b is novel in its 3' location to the A' site, and site3 lies upstream of A0 in a position analogous to the yeast U3-binding site. Here, in vivo analysis of mutated 5'ETS sequences shows that sequences 5' to the A' site are not needed for A' cleavage or SSU rRNA production. A' cleavage is linked to, but is not sufficient to trigger, downstream pre-SSU rRNA processing events. These events require an intact 11 nt sequence, 3'-adjacent to A', which directs efficient and accurate A' cleavage. Neither the A' nearby site1b nor the site3 U3-binding elements affect A' processing, yet each is required for A0 and A1 cleavage, and SSU rRNA production. The same U3 3' hinge bases evidently bind a core element, UGUu/gGGU, within site1a and site3; the U3-site1b interaction is less reliant on base-pairing than the U3-site3 interaction. As yeast U3 5' hinge bases pair to 5'ETS sequences, it is clear that distinct U3 hinge regions can interact at both novel and related 5'ETS sites to promote 3'-proximal 5'ETS processing events in diverse organisms. The T. brucei data fit a model wherein processing factors assemble at the 5'ETS site1a to affect A' cleavage and stabilize a U3-site1b complex, which may work in concert with the downstream U3-site3 complex to assist processing events leading to ribosomal SSU production.

Molecular characterization at the RNA and gene levels of U3 snoRNA from a unicellular green alga, Chlamydomonas reinhardtii

A U3 snoRNA gene isolated from a Chlamydomonas reinhardtii (CRE:) genomic library contains putative pol III-specific transcription signals similar to those of RNA polymerase III-specific small nuclear (sn)RNA genes of higher plants. The 222 nt long CRE: U3 snoRNA was immunoprecipitated by anti-gamma-mpppN antisera, but not by anti-m(2,2,7)G antibodies, supporting the notion that it is a RNA polymerase III transcript. Tagged CRE: U3 snoRNA gene constructs were expressed in CRE: cells. Results of chemical and enzymatic structure probing of CRE: U3 snoRNA in solution and of DMS modification of CRE: U3 snoRNA under in vivo conditions revealed that the two-hairpin structure of the 5'-domain that is found in solution is no longer detected under in vivo conditions. The observed differences can be explained by the formation of several base pair interactions with the 18S and 5'-ETS parts of the pre-rRNA. A model that involves five intermolecular helices is proposed.

Interaction of nucleolin with an evolutionarily conserved pre-ribosomal RNA sequence is required for the assembly of the primary processing complex

The first processing event of the precursor ribosomal RNA (pre-rRNA) takes place within the 5' external transcribed spacer. This primary processing requires conserved cis-acting RNA sequence downstream from the cleavage site and several nucleic acids (small nucleolar RNAs) and proteins trans-acting factors including nucleolin, a major nucleolar protein. The specific interaction of nucleolin with the pre-rRNA is required for processing in vitro. Xenopus laevis and hamster nucleolin interact with the same pre-rRNA site and stimulate the processing activity of a mouse cell extract. A highly conserved 11-nucleotide sequence located 5-6 nucleotides after the processing site is required for the interaction of nucleolin and processing. In vitro selection experiments with nucleolin have identified an RNA sequence that contains the UCGA motif present in the 11-nucleotide conserved sequence. The interaction of nucleolin with pre-rRNA is required for the formation of an active processing complex. Our findings demonstrate that nucleolin is a key factor for the assembly and maturation of pre-ribosomal ribonucleoparticles.

Identification of the transcription initiation site of the asexually expressed rRNA genes of the malaria parasite Plasmodium berghei

The start site of the A-type ribosomal RNA transcription units of the rodent malaria parasite, Plasmodium berghei, has been identified. The two A-type units cannot be distinguished within the transcription unit, yet exist as single copies on different chromosomes. Gene transcription initiates 820 bp upstream of the A-type small subunit (SSU) ribosomal gene and two major processing sites were mapped 610 and 611 nucleotides upstream of the SSU in the external transcribed spacer region. Surprisingly the nucleotide sequence of the DNA region containing the putative ribosomal promoter lacked repetitive DNA sequences typical of ribosomal promoters. This region was further analysed by computer using programs designed to reveal sequence-dependent structural features. Comparison of DNA curvature, duplex stability and pattern of twist angle variation revealed a striking degree of conservation between the ribosomal promoters from Plasmodium and other eukaryotes.

Nucleolin functions in the first step of ribosomal RNA processing

The first processing step of precursor ribosomal RNA (pre-rRNA) involves a cleavage within the 5' external transcribed spacer. This processing requires sequences downstream of the cleavage site which are perfectly conserved among human, mouse and Xenopus and also several small nucleolar RNAs (snoRNAs): U3, U14, U17 and E3. In this study, we show that nucleolin, one of the major RNA-binding proteins of the nucleolus, is involved in the early cleavage of pre-rRNA. Nucleolin interacts with the pre-rRNA substrate, and we demonstrate that this interaction is required for the processing reaction in vitro. Furthermore, we show that nucleolin interacts with the U3 snoRNP. Increased levels of nucleolin, in the presence of the U3 snoRNA, activate the processing activity of a S100 cell extract. Our results suggest that the interaction of nucleolin with the pre-rRNA substrate might be a limiting step in the primary processing reaction. Nucleolin is the first identified metazoan proteinaceous factor that interacts directly with the rRNA substrate and that is required for the processing reaction. Potential roles for nucleolin in the primary processing reaction and in ribosome biogenesis are discussed.

U3 snoRNA genes with and without intron in the Kluyveromyces genus: yeasts can accommodate great variations of the U3 snoRNA 3'-terminal domain

The U3 snoRNA coding sequences from the genomic DNAs of Kluyveromyces delphensis and four variants of the Kluyveromyces marxianus species were cloned by PCR amplification. Nucleotide sequence analysis of the amplification products revealed a unique U3 snoRNA gene sequence in all the strains studied, except for K. marxianus var. fragilis. The K. marxianus U3 genes were intronless, whereas an intron similar to those of the Saccharomyces cerevisiae U3 genes was found in K. delphensis. Hence, U3 genes with and without intron are found in yeasts of the Saccharomycetoideae subfamily. The secondary structure of the K. delphensis pre-U3 snoRNA and of the K. marxianus mature snoRNAs were studied experimentally. They revealed a strong conservation in yeasts of (1) the architecture of U3 snoRNA introns, (2) the 5'-terminal domain of the mature snoRNA, and (3) the protein-anchoring regions of the U3 snoRNA 3' domain. In contrast, stem-loop structures 2, 3, and 4 of the 3' domain showed great variations in size, sequence, and structure. Using a genetic test, we show that, in spite of these variations, the Kluyveromyces U3 snoRNAs are functional in S. cerevisiae. We also show that S. cerevisiae U3A snoRNAs lacking the stem-loop structure 2 or 4 are functional. Hence, U3 snoRNA function can accommodate great variations of the RNA 3'-terminal domain.

Analysis of nucleolar transcription and processing domains and pre-rRNA movements by in situ hybridization

We have examined the cytological localization of rRNA synthesis, transport, and processing events within the mammalian cell nucleolus by double-label fluorescent in situ hybridization analysis using probes for small selected segments of pre-rRNA, which have known half-lives. In particular, a probe for an extremely short-lived 5' region that is not found separate of the pre-rRNA identifies nascent transcripts within the nucleolus of an intact active cell, while other characterized probes identify molecules at different stages in the rRNA processing pathway. Through these studies, visualized by confocal and normal light microscopy, we (1) confirm that rDNA transcription occurs in small foci within nucleoli, (2) show that the nascent pre-rRNA transcripts and most likely also the rDNA templates are surprisingly extended in the nucleolus, (3) provide evidence that the 5' end of the nascent rRNA transcript moves more rapidly away from the template DNA than does the 3' end of the newly released transcript, and (4) demonstrate that the various subsequent rRNA processing steps occur sequentially further from the transcription site, with each early processing event taking place in a distinct nucleolar subdomain. These last three points are contrary to the generally accepted paradigms of nucleolar organization and function. Our findings also imply that the nucleolus is considerably more complex than the conventional view, inferred from electron micrographs, of only three kinds of regions - fibrillar centers, dense fibrillar components, and granular components - for the dense fibrillar component evidently consists of several functionally distinct sub-domains that correlate with different steps of ribosome biogenesis.

Early stages of pre-rRNA formation within the nucleolar ultrastructure of mouse cells studied by in situ hybridization with a 5'ETS leader probe

The first cleavage in the processing of the rRNA primary transcript in mammals occurs within the 5'-terminal region of the 5' external transcribed spacer (5'ETS), which makes the upstream portion of this spacer a selective marker of nascent transcripts. Moreover, short treatments with low doses of actinomycin D (AMD), which selectively suppress pre-rRNA synthesis and allow processing of preformed pre-rRNAs, result in the production of prematurely terminated transcripts essentially spanning the 5'ETS leader region. To gain further insight into the intranucleolar localization of early stages of preribosome formation we analyzed the distribution of this specific pre-rRNA segment by in situ hybridization at the ultrastructural level in AMD-treated or in control 3T3 mouse cells. In control cells, 5'ETS leader rRNA was detected at the border of the fibrillar centers and over the dense fibrillar component, in agreement with previous data suggesting that rRNA gene transcription takes place at the border of the fibrillar centers before a rapid transfer of the nascent trancript to the dense fibrillar component. Observation of cells subjected to a short treatment with low doses of AMD fully supports this conclusion, with the prematurely terminated 5'ETS leader-containing transcripts detected at the border of enlarged fibrillar centers. With prolonged periods of AMD treatment even the partial transcription of rRNA genes is blocked and fibrillar centers of typically segregated nucleoli show no positive signals with the 5'ETS leader probe. We also analyzed in parallel the intranucleolar distribution of U3 small nucleolar RNA, which is involved in 5'ETS processing, by hybridization with biotinylated antisense oligonucleotides. Distribution of U3 roughly paralleled that of 5'ETS leader rRNA in untreated cells. However, U3 RNA persisted in the dense fibrillar component of segregated nucleoli whatever the conditions of drug treatment, i.e., even after a thorough chase of the rRNA precursors from this nucleolar compartment.

Cytokinin induction of RNA polymerase I transcription in Arabidopsis thaliana

RNA polymerase I (pol I) transcribes the repeated genes that encode the precursor of 17-18, 5.8, and 25-28 S ribosomal RNA (rRNA). Pol I transcription is up-regulated in growing cells and down-regulated in quiescent cells, presumably reflecting the demand for ribosomes and protein synthesis. However, the signal transduction pathways responsible for pol I regulation are poorly understood. We tested the effects of exogenously applied plant hormones on promoter-dependent rRNA transcription in Arabidopsis thaliana. Gibberellic acid, abscisic acid, auxin, and ethylene had no detectable effect on rRNA transcription, but kinetin (a cytokinin) stimulated rRNA transcription within 1 h of treatment. Increased steady-state levels of accurately initiated rRNA transcripts, detected by S1 nuclease protection, were paralleled by increased levels of nascent rRNA transcripts in isolated nuclei. Therefore, the primary effect of cytokinin appears to be at the level of transcription initiation rather than rRNA stability. Pol I accounts for approximately 34% of total nuclear transcription in untreated plants and approximately 60% following cytokinin treatment. The specific responsiveness of pol I transcription to kinetin suggests that cytokinins may act as general regulators of protein synthetic capacity and growth status in plant cells.

Processing of mammalian rRNA precursors at the 3' end of 18S rRNA. Identification of cis-acting signals suggests the involvement of U13 small nucleolar RNA

Molecular mechanisms involved in the nucleolytic cleavage at the 18S rRNA/internal transcribed spacer 1 (ITS 1) junction, a late step of small-subunit pre-rRNA processing in vertebrates, remain largely unknown, mostly due to the lack of faithful in vitro assays. To identify the minimal cis-acting signals required for this reaction, we studied the processing of truncated human rRNA gene transcripts transiently expressed upon transfection of rRNA minigenes into cultured mouse cells. We observed that processing at this site was faithfully reproduced with transcripts containing only 60 nucleotides of 18S rRNA and the adjacent 103 nucleotides of ITS 1, but was abolished or severely altered by further shortening of either sequence. Remarkably, this minimal transcript contains, within its 18S rRNA part, long sequences complementary to both U20 and U13 small nucleolar RNAs (snoRNAs). The cis-acting elements essential for the reaction were studied further by site-directed mutagenesis. The U20 snoRNA complementary region in 18S rRNA was not required for faithful processing at the 18S rRNA/ITS 1 junction. Also, processing at this site was not appreciably altered by random substitution of proximal ITS 1 sequences (including the 5' terminal nucleotide) or of the terminal nucleotide of mature 18S rRNA. Substitutions in the four-nucleotide loop of the 18S rRNA 3'-terminal stem-loop, including the two adenosine residues substrates of dimethylation, did not alter appreciably the formation of the 18S rRNA 3' end, showing that the (methyl)2A1850.(methyl)2A1851 doublet was not required for processing at this site. Two highly conserved 18S rRNA elements acted as major cis-acting signals for processing at the 3' end, the CAUU sequence immediately preceding the 3'-terminal nucleotide and the 3' strand of the 3'-terminal 18S rRNA helix, complementary to U13 snoRNA. Compensatory mutations, restoring the potential for helix formation, but not U13 snoRNA complementarity, did not restitute the cleavage at the 3' end of 18S rRNA. This suggests that U13 snoRNA may be a trans-acting factor in the nucleolytic cleavage at the 3' end of 18S rRNA.

The yeast Hansenula wingei U3 snoRNA gene contains an intron and its coding sequence co-evolved with the 5' ETS region of the pre-ribosomal RNA

The 5' external transcribed spacer (ETS) region of the pre-rRNA in Saccharomyces cerevisiae contains a sequence with 10 bp of perfect complementarity to the U3 snoRNA. Base pairing between these sequences has been shown to be required for 18S rRNA synthesis, although interaction over the full 10 bp of complementarity is not required. We have identified the homologous sequence in the 5' ETS from the evolutionarily distant yeast Hansenula wingei; unexpectedly, this shows two sequence changes in the region predicted to base pair to U3. By PCR amplification and direct RNA sequencing, a single type of U3 snoRNA coding sequence was identified in H. wingei. As in the S. cerevisiae U3 snoRNA genes, it is interrupted by an intron with features characteristic of introns spliced in a spliceosome. Consequently, this unusual property is not restricted to the yeast genus Saccharomyces. The introns of the H. wingei and S. cerevisiae U3 genes show strong differences in length and sequence, but are located at the same position in the U3 sequence, immediately upstream of the phylogenetically conserved Box A region. The 3' domains of the H. wingei and S. cerevisiae U3 snoRNAs diverge strongly in primary sequence, but have very similar predicted secondary structures. The 5' domains, expected to play a direct role in pre-ribosomal RNA maturation, are more conserved. The sequence predicted to base pair to the pre-rRNA contains two nucleotide substitutions in H. wingei that restore 10 bp of perfect complementarity to the 5' ETS. This is a strong phylogenetic evidence for the importance of the U3/pre-rRNA interaction.

Processing of pre-ribosomal RNA in Saccharomyces cerevisiae

Post-transcriptional processing of precursor-ribosomal RNA comprises a complex pathway of endonucleolytic cleavages, exonucleolytic digestion and covalent modifications. The general order of the various processing steps is well conserved in eukaryotic cells, but the underlying mechanisms are largely unknown. Recent analysis of pre-rRNA processing, mainly in the yeast Saccharomyces cerevisiae, has significantly improved our understanding of this important cellular activity. Here we will review the data that have led to our current picture of yeast pre-rRNA processing.

Processing of eukaryotic pre-rRNA: the role of the transcribed spacers

The 17-18S, 5.8S, and 25-28S rRNA species of eukaryotic cells are produced by a series of nucleolytic reactions that liberate the mature rRNAs from the large primary precursor transcript synthesized by RNA polymerase 1. Whereas the order of the cleavage reactions has long been established, until recently little information was available on their molecular details, such as the nature of the proteins, including the nucleolytic enzymes, involved and the signals directing the processing machinery to the correct sites. This situation is now rapidly changing, in particular where yeast is concerned. The use of recently developed systems for in vivo mutational analysis of yeast rDNA has considerably enhanced our knowledge of cis-acting structural features within the pre-rRNA, in particular the transcribed spacer sequences, that are critical for correct and efficient removal of these spacers. The same systems also allow a link to be forged between trans-acting processing factors and these cis-acting elements. In this paper, we will focus predominantly on the nature and role of the cis-acting processing elements as identified in the transcribed spacer regions of Saccharomyces cerevisiae pre-rRNA.

Processing of truncated mouse or human rRNA transcribed from ribosomal minigenes transfected into mouse cells

The processing of pre-rRNA in eukaryotic cells involves a complex pattern of nucleolytic reactions taking place in preribosomes with the participation of several nonribosomal proteins and small nuclear RNAs. The mechanism of these reactions remains largely unknown, mainly because of the absence of faithful in vitro assays for most processing steps. We have developed a pre-rRNA processing system using the transient expression of ribosomal minigenes transfected into cultured mouse cells. Truncated mouse or human rRNA genes are faithfully transcribed under the control of mouse promoter and terminator signals. The fate of these transcripts is analyzed by the use of reporter sequences flanking the rRNA gene inserts. Both mouse and human transcripts, containing the 3' end of 18S rRNA-encoding DNA (rDNA), internal transcribed spacer (ITS) 1, 5.8S rDNA, ITS 2, and the 5' end of 28S rDNA, are processed predominantly to molecules coterminal with the natural mature rRNAs plus minor products corresponding to cleavages within ITS 1 and ITS 2. To delineate cis-acting signals in pre-rRNA processing, we studied series of more truncated human-mouse minigenes. A faithful processing at the 18S rRNA/ITS 1 junction can be observed with transcripts containing only the 60 3'-terminal nucleotides of 18S rRNA and the 533 proximal nucleotides of ITS 1. However, further truncation of 18S rRNA (to 8 nucleotides) or of ITS 1 (to 48 nucleotides) abolishes the cleavage of the transcript. Processing at the ITS 2/28S rRNA junction is observed with truncated transcripts lacking the 5.8S rRNA plus a major part of ITS 2 and containing only 502 nucleotides of 28S rRNA. However, further truncation of the 28S rRNA segment to 217 nucleotides abolishes processing. Minigene transcripts containing most internal sequences of either ITS 1 or ITS 2, but devoid of ITS/mature rRNA junctions, are not processed, suggesting that the cleavages in vivo within either ITS segment are dependent on the presence in cis of mature rRNA sequences. These results show that the major cis signals for pre-rRNA processing at the 18S rRNA/ITS 1 or the ITS2/28S rRNA junction involve solely a limited critical length of the respective mature rRNA and adjacent spacer sequences.

Processing of truncated mouse or human rRNA transcribed from ribosomal minigenes transfected into mouse cells

The processing of pre-rRNA in eukaryotic cells involves a complex pattern of nucleolytic reactions taking place in preribosomes with the participation of several nonribosomal proteins and small nuclear RNAs. The mechanism of these reactions remains largely unknown, mainly because of the absence of faithful in vitro assays for most processing steps. We have developed a pre-rRNA processing system using the transient expression of ribosomal minigenes transfected into cultured mouse cells. Truncated mouse or human rRNA genes are faithfully transcribed under the control of mouse promoter and terminator signals. The fate of these transcripts is analyzed by the use of reporter sequences flanking the rRNA gene inserts. Both mouse and human transcripts, containing the 3' end of 18S rRNA-encoding DNA (rDNA), internal transcribed spacer (ITS) 1, 5.8S rDNA, ITS 2, and the 5' end of 28S rDNA, are processed predominantly to molecules coterminal with the natural mature rRNAs plus minor products corresponding to cleavages within ITS 1 and ITS 2. To delineate cis-acting signals in pre-rRNA processing, we studied series of more truncated human-mouse minigenes. A faithful processing at the 18S rRNA/ITS 1 junction can be observed with transcripts containing only the 60 3'-terminal nucleotides of 18S rRNA and the 533 proximal nucleotides of ITS 1. However, further truncation of 18S rRNA (to 8 nucleotides) or of ITS 1 (to 48 nucleotides) abolishes the cleavage of the transcript. Processing at the ITS 2/28S rRNA junction is observed with truncated transcripts lacking the 5.8S rRNA plus a major part of ITS 2 and containing only 502 nucleotides of 28S rRNA. However, further truncation of the 28S rRNA segment to 217 nucleotides abolishes processing. Minigene transcripts containing most internal sequences of either ITS 1 or ITS 2, but devoid of ITS/mature rRNA junctions, are not processed, suggesting that the cleavages in vivo within either ITS segment are dependent on the presence in cis of mature rRNA sequences. These results show that the major cis signals for pre-rRNA processing at the 18S rRNA/ITS 1 or the ITS2/28S rRNA junction involve solely a limited critical length of the respective mature rRNA and adjacent spacer sequences.

Processing of eukaryotic ribosomal RNA

In summary, it can be argued that the understanding of eukaryotic rRNA processing is no less important than the understanding of mRNA maturation, since the capacity of a cell to carry out protein synthesis is controlled, in part, by the abundance of ribosomes. Processing of pre-rRNA is highly regulated, involving many cellular components acting either alone or as part of a complex. Some of these components are directly involved in the modification and cleavage of the precursor rRNA, while others direct the packaging of the rRNA into ribosome subunits. As is the case for pre-mRNA processing, snoRNPs are clearly involved in eukaryotic rRNA processing, and have been proposed to assemble with other proteins into at least one complex called a "processosome" (17), which carries out the ordered processing of the pre-rRNA and its assembly into ribosomes. The formation of a processing complex clearly makes possible the regulation required to coordinate the abundance of ribosomes with the physiological and developmental changes of a cell. It may be that eukaryotic rRNA processing is even more complex than pre-mRNA maturation, since pre-rRNA undergoes extensive nucleotide modification and is assembled into a complex structure called the ribosome. Undoubtedly, features of the eukaryotic rRNA-processing pathway have been conserved evolutionarily, and the genetic approach available in yeast research (6) should provide considerable knowledge that will be useful for other investigators working with higher eukaryotic systems. Interestingly, it was originally hoped that the extensive work and understanding of bacterial ribosome formation would provide a useful paradigm for the process in eukaryotes. However, although general features of ribosome structure and function are highly conserved between bacterial and eukaryotic systems, the basic strategy in ribosome biogenesis seems to be, for the most part, distinctly different. Thus, the detailed molecular mechanisms for rRNA processing in each kingdom will have to be independently deciphered in order to elucidate the features and regulation of this important process for cell survival.

Selection of a preribosomal RNA processing site by a nucleolar endoribonuclease involves formation of a stable complex

A nucleolar endoribonuclease from mouse Ehrlich ascites tumor cells, that has been implicated in the endonucleolytic cleavage of mouse precursor ribosomal RNA, specifically and stably binds an in vitro-derived rRNA transcript containing the +650 early processing site. The specificity of binding was demonstrated by mobility shift analysis, glycerol gradient velocity sedimentation analysis, and UV-crosslinking studies. Binding did not require Mg2+ and therefore was not dependent on cleavage; however, binding was dependent on the presence of the early +650 processing site since a pre-rRNA transcript with the +650 processing site deleted failed to compete in binding. A small nucleolar RNA component was not required for the formation of this stable complex or for the specific cleavage of a processing competent pre-rRNA transcript. UV crosslinking studies using 32P-labeled 5-azidouridine-substituted pre-rRNA with bound nucleolar endoribonuclease identified three closely sized polypeptides of approximately 50, approximately 48, and approximately 45 kDa, respectively, that specifically crosslinked to the processing competent rRNA transcript. These three polypeptides species were identified following ribonuclease digestion and electrophoresis on a SDS-polyacrylamide gel. An identical pattern of labeled polypeptides was also identified from gel mobility shift analysis where the specifically shifted material was U.V. crosslinked. The largest of these polypeptides corresponded to the estimated size of the nucleolar endoribonuclease, while the lower molecular weight species may represent partially proteolyzed enzyme. Overall, these results suggest that the unique specificity of the nucleolar endoribonuclease may, in part, be attributed to the formation of a stable complex at the +650 processing site for mouse preribosomal RNA, and that formation of this unique stable complex affords a means to specifically label the limited amount of available partially purified enzyme for sequence analysis.

A U3 small nuclear ribonucleoprotein-requiring processing event in the 5' external transcribed spacer of Xenopus precursor rRNA

A processing site has been identified within the 5' external transcribed spacer (ETS) of Xenopus laevis and X. borealis pre-RNAs, and this in vivo processing can be reproduced in vitro. It involves a stable and specific association of the pre-rRNA with factors in the cell extract, including at least four RNA-contacting polypeptides, yielding a distinct complex that sediments at 20S. Processing also requires the U3 small nuclear RNA. This processing, at residue +105 of the 713-nucleotide X. laevis 5' ETS, is highly reminiscent of the initial processing cleavage of mouse pre-rRNA within its 3.5-kb 5' ETS, previously thought to be mammal specific. The frog and mouse processing signals share a short essential sequence motif, and mouse factors can faithfully process the frog pre-rRNA. This conservation suggests that this 5' ETS processing site serves an evolutionarily selective function.

A U3 small nuclear ribonucleoprotein-requiring processing event in the 5' external transcribed spacer of Xenopus precursor rRNA

A processing site has been identified within the 5' external transcribed spacer (ETS) of Xenopus laevis and X. borealis pre-RNAs, and this in vivo processing can be reproduced in vitro. It involves a stable and specific association of the pre-rRNA with factors in the cell extract, including at least four RNA-contacting polypeptides, yielding a distinct complex that sediments at 20S. Processing also requires the U3 small nuclear RNA. This processing, at residue +105 of the 713-nucleotide X. laevis 5' ETS, is highly reminiscent of the initial processing cleavage of mouse pre-rRNA within its 3.5-kb 5' ETS, previously thought to be mammal specific. The frog and mouse processing signals share a short essential sequence motif, and mouse factors can faithfully process the frog pre-rRNA. This conservation suggests that this 5' ETS processing site serves an evolutionarily selective function.

The terminal balls characteristic of eukaryotic rRNA transcription units in chromatin spreads are rRNA processing complexes

When spread chromatin is visualized by electron microscopy, active rRNA genes have a characteristic Christmas tree appearance: From a DNA "trunk" extend closely packed "branches" of nascent transcripts whose ends are decorated with terminal "balls." These terminal balls have been known for more than two decades, are shown in most biology textbooks, and are reported in hundreds of papers, yet their nature has remained elusive. Here, we show that a rRNA-processing signal in the 5'-external transcribed spacer (ETS) of the Xenopus laevis ribosomal primary transcript forms a large, processing-related complex with factors of the Xenopus oocyte, analogous to 5' ETS processing complexes found in other vertebrate cell types. Using mutant rRNA genes, we find that the same rRNA residues are required for this biochemically defined complex formation and for terminal ball formation, analyzed electron microscopically after injection of these cloned genes into Xenopus oocytes. This, plus other presented evidence, implies that rRNA terminal balls in Xenopus, and by inference, also in the multitude of other species where they have been observed, are the ultrastructural visualization of an evolutionarily conserved 5' ETS processing complex that forms on the nascent rRNA.

Association with terminal exons in pre-mRNAs: a new role for the U1 snRNP?

Psoralen cross-linking experiments in HeLa cell nuclear extracts have revealed the binding of U1 snRNA to substrates containing the SV40 late and adenovirus L3 polyadenylation signals. The sites of U1 cross-linking to the substrates map different distances upstream of the AAUAAA sequence to regions with limited complementarity to the 5' end of U1 snRNA. U1 cross-linking to the same site in the SV40 late pre-mRNA is enhanced by the addition of an upstream 3' splice site, which also enhances polyadenylation. Examination of different nuclear extracts reveals a correlation between U1 cross-linking and the coupling of splicing and polyadenylation, suggesting that the U1 snRNP participates in the coordination of these two RNA-processing events. Mutational analyses demonstrate that U1/substrate association cannot be too strong for coupling to occur and suggest that the U1 snRNP plays a similar role in recognition of internal and 3' terminal exons. Possible mechanisms for communication between the splicing and polyadenylation machineries are discussed, as well as how interaction of the U1 snRNP with 3' terminal exons might contribute to mRNA export.

A new interaction between the mouse 5' external transcribed spacer of pre-rRNA and U3 snRNA detected by psoralen crosslinking

The first cleavage in mammalian pre-rRNA processing occurs within the 5' external transcribed spacer (ETS). We have recently shown that the U3 snRNP is required for this cleavage reaction, binds to the rRNA precursor, and remains complexed with the downstream processing product after the reaction has been completed (1). Using psoralen crosslinking in mouse cell extract we have detected a new interaction between U3 RNA and the mouse ETS processing substrate and its processed product. The crosslinked sites on both U3 and ETS RNAs have been mapped by RNase H cleavage and primer extension analyses. The crosslinked sites in U3 RNA map to C5, U6, and U8. U8 lies within and C5 and U6 are adjacent to an evolutionarily conserved U3 sequence termed box A'. In the ETS the crosslinked sites are U1012 and U1013, 362 nucleotides downstream from the processing site. Although the crosslinked site is dispensable for the primary processing reaction in vitro, a short conserved sequence just 3' to the cleavage site (nucleotides 650-668) is absolutely required for crosslink formation. We conclude that the interaction between U3 RNA and the 5' ETS detected by psoralen crosslinking may play a role in subsequent step(s) of pre-rRNA processing.

Nuclear RNase MRP processes RNA at multiple discrete sites: interaction with an upstream G box is required for subsequent downstream cleavages

RNase MRP is a site-specific endoribonuclease that processes primer RNA from the leading-strand origin of mammalian mitochondrial DNA replication. It is present in active form as isolated from the nucleus, suggesting a bipartite cellular location and function. The relatively high abundance of nucleus-localized RNase MRP has permitted its purification to near homogeneity and, in turn, has led to the identification of protein components of this ribonucleoprotein. Analysis of the mode of RNA cleavage by nuclear RNase MRP revealed the surprising and unprecedented ability of the endonuclease to process RNA at multiple discrete locations. Substrate cleavage is dependent on the presence of a previously described G-rich sequence element adjacent to the primary site of RNA processing. Downstream cleavage occur in a distance- and sequence-specific manner.

Expression of mouse and frog rRNA genes: transcription and processing

This article summarizes a number of lines of investigation of rRNA gene expression that are ongoing in the laboratory. These studies focus on mouse and frog, two distant vertebrate species. One major conclusion is that the basic properties of rRNA gene expression appear remarkably well conserved in evolution, with only relatively minor perturbations between frog and mouse, contrary to the common interpretation of the species-selectively between mouse and human rDNA transcription (e.g., 1). This is true both for the process of rDNA transcription and for the subsequent rRNA processing event.

Sequence organization and RNA structural motifs directing the mouse primary rRNA-processing event

The first processing step in the maturation of mouse precursor rRNA involves cleavage at nucleotide ca. +650, at the 5' border of a 200-nucleotide region that is conserved across mammals and contains the sequences that direct the processing. To identify the relevant sequence elements, we used rRNAs with small internal mutations and short pre-rRNA substrates. Much of the region can be mutated without appreciable effect, but nucleotides +655 to +666 appear to be absolutely required and short segments surrounding +750 and +810 markedly stimulate processing. The minimal processing signal corresponds to rRNA nucleotides +645 to +672. Formation of a ribonucleoprotein complex of retarded electrophoretic mobility is evidently necessary but not sufficient for processing. Computer-assisted analysis suggested a phylogenetic- and mutant-supported secondary structure in which the minimal processing signal forms a stem with the +655 region in the loop, and there is a separate branched duplex containing the downstream stimulatory sequences. Use of antisense RNA, in trans and in cis, to sequester the +655 region in a duplex supported the hypothesis that this critical region was needed in a single-stranded conformation for processing and for specific complex formation.

Secondary structure of the 5' external transcribed spacer of vertebrate pre-rRNA. Presence of phylogenetically conserved features

Eucaryotic pre-rRNA spacers are evolutionarily highly variable in sequence and size, and are markedly expanded in vertebrates, particularly in mammals. The longest mammalian spacer by far is the 3.5-4-kb 5' external transcribed spacer (5'-ETS), which is excised in two steps. We present a folding model for the entire mammalian 5'-ETS, derived from comparative analyses and thermodynamic predictions for mouse, rat and human sequences, which should prove helpful in identifying cis-acting processing signals, particularly those involved in its early internal cleavage, for which U3 RNA is an essential factor. Although the rodent and primate sequences have extensively diverged, a series of relatively conserved sequence tracts can nevertheless be identified: they participate in base-pairing, preserved through the occurrence of compensatory base changes, which delineate four independent domains of secondary structure. The first domain is located entirely upstream from the site of internal cleavage. The second domain, immediately downstream from this cleavage site, encompasses most of the region required for faithful and efficient in vitro processing at this site. Phylogenetically supported conserved structures also define two other independent domains, encompassing most of the 5'-ETS length, with the presence of giant hairpins (extending from the conserved core elements) which exhibit both some analogous features and substantial differences between man and mouse. The comparative analysis was extended to the two other vertebrate sequences available so far, amphibians Xenopus laevis and Xenopus borealis. The amphibian folding model, supported by comparative evidence between these two species, displays some features in common with the mammalian model, with a similar organization into four separate domains of secondary structure, suggesting the functional relevance of these structures in the process of ribosome formation.

Sequence organization and RNA structural motifs directing the mouse primary rRNA-processing event

The first processing step in the maturation of mouse precursor rRNA involves cleavage at nucleotide ca. +650, at the 5' border of a 200-nucleotide region that is conserved across mammals and contains the sequences that direct the processing. To identify the relevant sequence elements, we used rRNAs with small internal mutations and short pre-rRNA substrates. Much of the region can be mutated without appreciable effect, but nucleotides +655 to +666 appear to be absolutely required and short segments surrounding +750 and +810 markedly stimulate processing. The minimal processing signal corresponds to rRNA nucleotides +645 to +672. Formation of a ribonucleoprotein complex of retarded electrophoretic mobility is evidently necessary but not sufficient for processing. Computer-assisted analysis suggested a phylogenetic- and mutant-supported secondary structure in which the minimal processing signal forms a stem with the +655 region in the loop, and there is a separate branched duplex containing the downstream stimulatory sequences. Use of antisense RNA, in trans and in cis, to sequester the +655 region in a duplex supported the hypothesis that this critical region was needed in a single-stranded conformation for processing and for specific complex formation.

The first pre-rRNA-processing event occurs in a large complex: analysis by gel retardation, sedimentation, and UV cross-linking

The first processing event that mouse pre-rRNA undergoes occurs within the external transcribed spacer and is efficiently reproduced in vitro. Analysis with nondenaturing polyacrylamide gels revealed the formation of heparin-resistant complexes of retarded electrophoretic mobility on the substrate rRNA. The specificity of these complexes was demonstrated by their elimination due to competition with processing-competent, but not with processing-incompetent, rRNAs. Furthermore, complex formation, like the processing cleavage, required only 28 nucleotides of rRNA sequence adjacent to the processing site but was stimulated by additional downstream conserved sequences. These processing complexes formed in a time-dependent manner, and once assembled, they were stable to challenge by competitor rRNA and remained on the processed rRNA. Their sedimentation coefficient was approximately 20S. UV cross-linking studies with 4-thiouridine-substituted rRNA have identified six polypeptides, 52 to 250 kilodaltons, that are specifically bound to the rRNA processing substrate.

The first pre-rRNA-processing event occurs in a large complex: analysis by gel retardation, sedimentation, and UV cross-linking

The first processing event that mouse pre-rRNA undergoes occurs within the external transcribed spacer and is efficiently reproduced in vitro. Analysis with nondenaturing polyacrylamide gels revealed the formation of heparin-resistant complexes of retarded electrophoretic mobility on the substrate rRNA. The specificity of these complexes was demonstrated by their elimination due to competition with processing-competent, but not with processing-incompetent, rRNAs. Furthermore, complex formation, like the processing cleavage, required only 28 nucleotides of rRNA sequence adjacent to the processing site but was stimulated by additional downstream conserved sequences. These processing complexes formed in a time-dependent manner, and once assembled, they were stable to challenge by competitor rRNA and remained on the processed rRNA. Their sedimentation coefficient was approximately 20S. UV cross-linking studies with 4-thiouridine-substituted rRNA have identified six polypeptides, 52 to 250 kilodaltons, that are specifically bound to the rRNA processing substrate.

Transcription of tomato ribosomal DNA and the organization of the intergenic spacer

The organization of the intergenic spacer of a 9.04 kb tomato ribosomal RNA gene (rDNA) was determined. The 3258 bp spacer contains two major repeat elements enclosing a region which includes 351 bp of an 81.8% A --T rich sequence. A block of nine 53 bp repeats begins 388 bp downstream from the 3' end of the 25S rRNA. The A--T rich domain is followed by a block of six 141 bp repeats terminating 818 bp upstream from the 5' end of the 18S rRNA. Major pre-rRNAs of 7.6 and 6.5 kb were observed by Northern hybridization analysis. The 5' termini of these RNAs were identified through combined S1 nuclease and primer extension analyses. The 7.6 kb RNA is likely to be the primary transcript; its 5' terminus lies within a sequence motif. TATA(R)TA(N)GGG, conserved at the termini of transcripts mapped in three other plant species. The 6.5 kb RNA is interpreted as a 5' end processed transcript derived from the 7.6 kb RNA. Comparative analysis of transcribed sequences revealed a 25 bp domain of the intergenic spacer which is relatively conserved among five plant species. The conservation of spacer sequences in plants is in contrast to the extensive sequence divergence of the intergenic spacer in other non-plant systems and suggests a conserved function directed by these sequences.

Accurate processing of human pre-rRNA in vitro

We report here that the mature 5' terminus of human 18S rRNA is generated in vitro by a two-step processing reaction. In the first step, SP6 transcripts were specifically cleaved in HeLa cell nucleolar extract at three positions near the external transcribed spacer (ETS)-18S boundary. Of these cleavage sites, two were major and the other was minor. RNase T1 fingerprint and secondary nuclease analyses placed the two major cleavage sites 3 and 8 bases upstream from the mature 5' end of 18S rRNA and the minor cleavage site 1 base into the 18S sequence. All three cleavages yielded 5'-hydroxyl, 2'-3'-cyclic phosphate termini and were 5' of adenosine residues in the sequence UACCU, which was repeated three times near the ETS-18S boundary. In the second step, the initial cleavage product containing 3 bases of ETS was converted to an RNA with a 5' terminus identical to that of mature 18S RNA by an activity found in HeLa cell cytoplasmic extracts.

Accurate processing of human pre-rRNA in vitro

We report here that the mature 5' terminus of human 18S rRNA is generated in vitro by a two-step processing reaction. In the first step, SP6 transcripts were specifically cleaved in HeLa cell nucleolar extract at three positions near the external transcribed spacer (ETS)-18S boundary. Of these cleavage sites, two were major and the other was minor. RNase T1 fingerprint and secondary nuclease analyses placed the two major cleavage sites 3 and 8 bases upstream from the mature 5' end of 18S rRNA and the minor cleavage site 1 base into the 18S sequence. All three cleavages yielded 5'-hydroxyl, 2'-3'-cyclic phosphate termini and were 5' of adenosine residues in the sequence UACCU, which was repeated three times near the ETS-18S boundary. In the second step, the initial cleavage product containing 3 bases of ETS was converted to an RNA with a 5' terminus identical to that of mature 18S RNA by an activity found in HeLa cell cytoplasmic extracts.

Analysis of pre-rRNAs in heat-shocked HeLa cells allows identification of the upstream termination site of human polymerase I transcription

Human rRNA precursor from normal or stressed HeLa cells were studied by S1 nuclease mapping of unlabeled RNA and by antisense RNase mapping of RNA from cells that had been labeled in vivo with [32P]PO4. Heating cells to 43 degrees C decreased the amount of newly synthesized rRNA to less than 5% of the control level and led to greater than 95% inhibition of transcription termination at a region 355 to 362 nucleotides downstream of the 3' end of 28S rRNA, with readthrough continuing into the next transcription unit. Heating of cells to 42 degrees C led to 60% inhibition of termination at this site; 50% of transcripts that extended into the nontranscribed spacer ended in a region 200 to 210 nucleotides upstream of the polymerase I (Pol I) initiation site. This is presumed to be the human upstream transcription termination site because of the absence of RNAs with a 5' end corresponding to this region, the location relative to the Pol I initiation site (which is similar to the location of upstream terminators in other species), and the fact that it is 15 to 25 nucleotides upstream of the sequence GGGTTGACC, which has an 8-of-9 base identity with the sequence 3' of the downstream termination site. Surprisingly, treatment of cells with sodium arsenite, which also leads to the induction of a stress response, did not inhibit termination. Pol I initiation was decreased to the same extent as termination, which lends support to the hypothesis that termination and initiation are coupled. Although termination was almost completely inhibited at 43 degrees C, the majority of the recently synthesized rRNAs were processed to have the correct 3' end of 28S. This finding suggests that 3'-end formation can involve an endonucleolytic cut and is not solely dependent on exonucleolytic trimming of correctly terminated rRNAs.

U3 small nuclear RNA can be psoralen-cross-linked in vivo to the 5' external transcribed spacer of pre-ribosomal-RNA

U3 small nuclear RNA is hydrogen-bonded to high molecular weight nucleolar RNA and can be isolated from greater than 60S pre-ribosomal ribonucleoprotein particles, suggesting that it is involved in processing of ribosomal RNA precursors (pre-rRNA) or in ribosome biogenesis. Here we have used in vivo psoralen cross-linking to identify the region of pre-rRNA interacting with U3 RNA. Quantitative hybridization selection/depletion experiments with clones of rRNA-encoding DNA (rDNA) and cross-linked nuclear RNA showed that all of the cross-linked U3 RNA was associated with a region that includes the external transcribed spacer (ETS) at the 5' end of the human rRNA precursor. To further identify the site of interaction within the approximately 3.7-kilobase ETS, Southern blots of rDNA clones were sandwich-hybridized with cross-linked RNA and then probed for cross-linked U3 RNA. These experiments showed that U3 RNA was cross-linked to a 258-base sequence between nucleotides +438 and +695, just downstream of the ETS early cleavage site (+414). The localization of U3 to this region of the rRNA precursor was not expected from previous models for a base-paired U3-rRNA interaction and suggests that U3 plays a role in the initial pre-rRNA processing event.

Structure of the 5'-external transcribed spacer of the human ribosomal RNA gene

We report the complete nucleotide sequence of the 3627 bp long 5'-external transcribed spacer (ETS) of a human ribosomal RNA gene. This sequence exhibits only very limited homologies with its mouse counterpart, the only other mammalian specimen analyzed so far. It has very peculiar compositional characteristics, with a highly biased base content (very rich in G + C, very poor in A) and also some very strong dinucleotide preferences. Interestingly, these specific features are shared by the mouse sequence, despite the extensive sequence divergence, and also apply to the other transcribed spacers of mammals indicating that a common and strong structural constraint is exerted on all these regions of the ribosomal gene. An outstanding secondary structure can be formed within the human ETS RNA, which could have a significant role in preribosome assembly.