The primary function of the nucleolus is to biosynthesize ribosomes, both the large and small subunits. The nucleolus is the most conspicuous organelle within the nucleus. Because of its prominence, light microscopists have studied the nucleolus for more than 200 years. Fontana described the nucleolus as early as 1774, and there was a review of its literature in 1898 by Montgomery. Although much has been learned, the nucleolus is still imperfectly understood.
1. Overview
Assembly of large and small ribosomal subunits requires 18S, 5.8S, 28S, and 5S ribosomal RNA molecules (rRNA) and approximately 85 structural ribosomal proteins. The 18S, 5.8S, and 28S rRNA, but not 5S, are transcribed within the nucleolus by RNA polymerase I as parts of a large precursor transcript, pre-rRNA (47S in mammalian cells, 42S in Xenopus, and 35S in yeast) (see Ribosomes). RNA spacer sequences reside in front of the 18S region, between the 18S and 5.8S regions, and between the 5.8S and 28S regions. The pre-rRNA is processed within the nucleolus by several endonuclease-catalyzed cleavages to yield mature 18S, 5.8S, and 28S rRNA. Although still imperfectly understood, these cleavage reactions require several nucleolar-specific proteins and small nucleolar RNAs (see later). Processing of the 45 S pre-rRNA occurs concomitantly with ribosomal subunit assembly. The 5 S rRNA is transcribed by RNA polymerase III from multiple copies of a gene that resides outside the nucleolus. After biosynthesis, the 5S RNA must migrate through the nucleoplasm to the nucleolus. Most of the ribosomal proteins translocate into the nucleus from their site of synthesis in the cytoplasm. Then they associate with the maturing rRNA in the nucleolus during assembly of the ribosomal subunits. The small ribosomal subunit consists of 18S rRNA and small subunit-specific ribosomal proteins (S-proteins), and the large ribosomal subunit consists of 5.8S and 28S rRNA, the incoming 5S rRNA, and large subunit-specific ribosomal proteins (L proteins) (1). Small and large subunits move from the nucleoli to the nuclear envelope through a nuclear pore complex and into the cytoplasm. The subunits unite to form intact ribosomes once they associate with messenger RNA to participate in synthesizing new protein (see Translation).
The number of nucleoli per nucleus varies, depending on the cell type. Yeast cells, for example, contain one nucleolus that is relatively large compared with the volume of the nucleus, and most metazoan cells contain one or a few nucleoli. On the other hand, amphibian oocytes contain hundreds of nucleoli per nucleus because of selective rDNA amplification during the early pachytene stage of meiosis I (2). The number of nucleoli also varies during the cell cycle. For example, nucleoli disassemble at the onset of mitosis, and they reappear beginning in the telophase by a process called nucleologenesis. In the late telophase/early interphase, a new cell displays several small nucleoli. The number reflects the number of nucleolar organizers that are characteristic of the species. As the new cells mature through the interphase, their nucleoli fuse to form larger nucleoli, and the number of nucleoli actually declines.
1.1. Nucleolar Ultrastructure
The interphase nucleolus consist of three ultrastructural subdivisions: the fibrillar center (FC), the surrounding dense fibrillar regions (DFR), and the peripheral granular regions (GR) (1, 3). The DFR are darker than the FC when stained with uranyl acetate and viewed by electron microscopy. As the name implies, the GR consist of small granules believed to be the still immature ribosomal subunits (see later). The overall nucleolar morphology varies in different cell types because of differences in the relative sizes and distributions of these three subdivisions (3). For example, nucleoli within metabolically active cells contain DFR that are cast into nucleolonemae, which appear in the electron microscope as darkly-stained, anastomosing thread-like structures. Each thread is surrounded by lightly staining material. Conversely, nucleoli in a few cell types constitute a second morphological class and appear compact, with uniform RNA-containing structures. The third classification includes ring-shaped nucleoli in which RNA-containing structures form a peripheral ring around chromatin-like fibrils (3). Nucleolar morphology is influenced by the particular physiology of a given cell type. For example, the multiple nucleoli within transcriptionally active oocytes from Xenopus (stage IV) show pronounced nucleolonemae, but greatly reduced GR, perhaps because of rapid ribosomal subunit biosynthesis and export from the nucleus. Conversely, nucleoli in mature and relatively quiescent stage VI oocytes display collapsed DFR (eg, no anastomosing nucleolonemae) and very prominent GR (4).
The FCs clearly contain the DNA encoding the rRNA, and they may well be the interphase homologue of mitotic nucleolar organizers, the genetic loci that contain tandemly repeated ribosomal genes (see later). The precise function of the FC is controversial. RNA polymerase I (5, 6) and topoisomerase I (7) have been detected in the FC, thus suggesting a potential for transcription. Using selective in situ hybridization and immunolabeling techniques to detect both rDNA and rRNA within the FC, Thiry (8, 9) argued that transcription must occur within these centers. Hozak et al. (10), however, proposed an alternative model in which active transcription units radiate from condensed DNA within the FC. In this latter model, the transcription units occupy the borders between the FC and the DFR. Regardless of the precise site of transcription, the dense fibrillar regions are the sites for early pre-rRNA processing and initial ribosomal subunit assembly. This assembly is a vectoral process, and maturation of the large and small ribosomal subunits continues on into the GR. Certain ribosomal proteins associate early in the DFR, and others associate in the GR. Finally, ribosomal subunits pass out of the nucleus into the cytoplasm, where final assembly with a few more ribosomal proteins takes place before they are competent to function in mRNA translation (1).
1.2. Ribosomal RNA Gene Organization and Expression
Genes that encode the large 45S pre-rRNA are tandemly repeated head-to-tail within genomes of most, if not all, eukaryotic organisms. Intergenic spacer (IGS) DNA sequences separate the individual genes. There are approximately 200 copies of ribosomal genes per haploid genome in humans, 100 in mouse, 500 to 600 in Xenopus laevis, 150 in Drosophila melanogaster, and 140 in the budding yeast Saccharomyces cerevisiae (1). Heitz (11, 12) and McClintock (13) demonstrated that the secondary constrictions in mitotic chromosomes are the genetic loci that initiate nucleolar formation as cells leave mitosis and enter the interphase. McClintock actually coined the term "nucleolar organizer" to describe these secondary constrictions. Now we know that the tandemly repeated pre-rRNA genes reside at these loci and that transcription of these genes initiates formation of nucleoli during the late telophase (see Nucleolar Organizer). Mutations that map to the nucleolar organizers reduce the number of rRNA genes or eliminate nucleoli altogether. For example, the bobbed (bb) mutation on the X-chromosomein Drosophila is a partial deletion of the tandem rDNA repeats. Flies homozygous or hemizygous for the mutation display pleiotropic effects of slow development, reduced fertility, and shortened (bobbed) bristles. In Xenopus, the anucleate mutation Onu fails to synthesize rRNA because it lacks ribosomal RNA genes. Finally, McClintock (13) described a chromosomal deletion in maize that eliminates the nucleolar organizer and also the ability to form nucleoli.
One of the most informative techniques for elucidating nucleolar gene ultrastructure is the spreading of nucleolar chromatin in solutions with low salt and high pH for electron microscopic examination (14-16). These so-called "Miller spreads" show that nucleolar chromatin consists of nascent rRNP fibrils (rRNA associated with protein) which project from the tandemly repeated ribosomal genes. RNA polymerase I complexes pack the active ribosomal RNA genes, and the RNP fibrils remain attached to these complexes during preparation. As a result, the RNP fibrils display a gradient in length. The shortest fibrils are closest to the transcription initiation site, and the longest fibrils approach transcription termination. The gradient of RNP fibrils looks like a "Christmas tree" where the RNA polymerase I-coated gene is the tree trunk that supports the RNP branches. Interestingly, the 5′-end of each RNP fibril contains a relatively large RNP particle (see later for function). These particles have been called the "ornaments" on the Christmas trees.
The organization of a typical rRNA gene in higher eukaryotes is as follows beginning at the transcription start site: the 5′-external transcribed spacer (5′-ETS) region; the 18S rRNA region; the first internal transcribed spacer (ITS1); the 5.8S rRNA region; the second internal transcribed region (ITS2); the 28S rRNA region; and finally a short 3′-external transcribed spacer (3′-ETS). The Xenopus rDNA gene is described here because it is well characterized. The nascent transcript encoded by the Xenopus genes is 7,625 nucleotides long and has a sedimentation coefficient of 40S. Individual genes are separated by intergenic spacers (IGS) of variable length. For the vast majority of rRNA genes in Xenopus, transcription initiates at the promoter just upstream of the ETS and stops within the IGS at a site 235 bp downstream of the 28S region at the site called T2, yielding a 40S pre-rRNA transcript. The T1 site is at the very end of the 28S region, and it arises from initial posttranscriptional trimming of the 3′-ETS back to this site (17). The intergenic spacers in Xenopus contain several duplicated promoters and enhancers (18). Occasionally, transcription initiates at a distant upstream IGS promoter, reads through the other downstream promoters and enhancers (thus silencing these promoters), and then terminates at T3, a site approximately 60 bp upstream of the next rDNA proximal gene promoter (eg, 215 bp upstream of the transcription start site of the next gene). This secondary transcription unit is fully contained within the IGS. IGS transcription may positively enhance transcription of this next gene by terminating at T3 and then handing the polymerase complex over to the adjacent downstream promoter (19, 20). Although the actual site of transcription termination has been debated, electron microscopic examination verifies that the vast majority of rDNA transcription units terminate transcription at the T2 site, at least in Xenopus (21).
Despite length differences in regions that encode either the mature rRNA themselves or the transcribed spacers, the overall genetic organization is similar in other eukaryotes (1). Many species of Tetrahymena, however, contain an intron within the 28S region. This is a group I intron originally characterized as self-splicing (22; see Catalytic RNA). Many of the Drosophila rRNA genes also contain one of two types of introns in the 28 S region. One type (a transposable element) appears only in 50% of the X-chromosome ribosomal genes. The other type appears in the ribosomal genes of both the X- and Y-chromosomes but at a frequency of 15%. Yeast has the same overall ribosomal gene structure as the higher eukaryotic cells, except that the primary transcript is shorter than in vertebrates (35S versus 40 to 45S). Yeast produces a 25S mature ribosomal RNA instead of a 28S transcript. Finally, the intergenic sequences in S. cerevisiae and discoideum contain 5S genes that are transcribed by RNA polymerase III but in the direction opposite to that of the main rDNA repeats.
Besides RNA polymerase I, several transcription factors (TF) are required for efficient rDNA expression (19). The upstream binding factor (UBF, a protein of 97 kDa in humans) contains a triple a-helical domain characteristic of the HMG proteins in chromatin. On its own, UBF binds promiscuously to both rDNA enhancers and promoters. UBF forms a stable complex with another factor, called SL1 in humans and rats, which then selectively binds the rDNA promoter. Transcription is initiated when an active RNA polymerase I complex recognizes the UBF-SL1 complex situated on the proximal promoter. Another transcription initiation factor, called TIF-IC in mouse cells, regulates the ability of polymerase I to recognize the UBF/SL1 complex. Topoisomerase I is relatively abundant within nucleoli. Its primary function is thought to aid RNA polymerase I transcription by relieving torsional stress generated within the DNA by the process of transcription. Low concentrations of actinomycin D (0.05 to 25 |ig/ml) selectively inhibit RNA polymerase I transcription and strip the nascent RNP fibrils from the transcription units. This has been used to study nucleolar transcription, structure, and function (23).
Multiple copy, tandemly repeated genes encode the 5 S ribosomal RNA that assemble within the large ribosomal subunit. As mentioned above, the 5 S genes in Saccharomyces, Dictyostelium, and a few other organisms are closely linked to the rRNA genes within the same repeating units. The 5S genes in most other organisms are located on other chromosomes without a nucleolar organizer. In addition, the clusters of 5S genes are scattered between several loci in many organisms (1). The number of 5S genes within an organism usually exceeds the number of 45S rRNA genes. Xenopus, for example, has over 20,000 5S genes per haploid genome. The 5S genes are transcribed by RNA polymerase III and are regulated by transcription factors IIIA, IIIB, and IIIC. TFIIIA binds an intergenic control region of the 5S genes to direct the binding of TFIIIC. The 5S genes are small, 120 bp in total length. Therefore, the bound TFIIIA/C complex directs TFIIIB to bind just upstream of the transcription start site. RNA polymerase III finally associates with TFIIIB to initiate transcription. Interestingly, TFIIIB consists of three proteins, one of which is the TATA binding protein found in RNA polymerase II transcription initiation complexes. The 5S transcripts follow a rather circuitous route to the nucleoli. Monomeric ribosomal protein L5 binds 5S rRNA in the nucleus, and together they migrate out to the cytoplasm. They reenter the nucleus and then associate with the nucleolus, where they join in ribosomal assembly (24).
1.3. Nonribosomal Nucleolar Proteins
Besides RNA polymerase I, UBF, SL1, and topoisomerase I, which regulate 45 S rRNA gene transcription, many other nonribosomal nucleolar proteins are involved with pre-rRNA processing (25).
1.3.1. Fibrillarin
To date vertebrate fibrillarin (Nop1 in S. cerevisiae) is the best characterized in terms of function. Vertebrate fibrillarin has a molecular mass of 34 kDa. It localizes mostly to the DFR, either to discrete foci within the DFR or to rings that surround the FC. This sublocalization suggests functional compartmentalization of fibrillarin within the DFR. Fibrillarin is conserved throughout eukaryotes. The amino-terminal portion contains a glycine/arginine-rich domain (called a GAR or RGG domain). Nearly half of these arginine residues are dimethylated, a posttranslational modification reserved for many nuclear RNA-binding proteins that contain RGG domains (26). Fibrillarin also contains a consensus RNA-binding domain (CS-RBD) in its center (27). Fibrillarin closely associates with many small nucleolar RNA (snoRNA) involved in either cleavage or base modification of the pre-rRNA. The best characterized fibrillarin association is with snoRNA U3 to form the RNP complex necessary for the first cleavage event within the ETS region of the pre-rRNA (see later).
1.3.2. Nucleolin
Vertebrate nucleolin is a phosphoprotein of molecular mass 90-110 kDa that is abundant in rapidly dividing somatic cells (28, 29) and in developing amphibian oocytes (30), where rates of ribosomal production are maximal. Although nucleolin shuttles between nucleoli and the cytoplasm (31), the greatest steady state concentration of the protein remains within the nucleolar DFR. Lesser amounts are found in the GR, and very little if any nucleolin resides in the FC. Vertebrate nucleolin consists of modular domains (32). The amino-terminal third contains alternating acidic domains (containing glutamic acid, aspartic acid, and phosphorylated serine residues exclusively) and basic domains (containing lysine residues). The serines within the acidic regions are phosphorylated by casein kinase type II enzymes during the interphase (see Phosphorylation). The basic domains, on the other hand, contain mitotic p34cdc2/cyclin B (MPF, maturation promoting factor) phosphorylation sites whose sequence is -Thr-Pro-(Ala or Gly)-Lys-Lys- (ie, -TP A/GKK-). Differential phosphorylation within the amino-terminal domain of nucleolin may regulate nucleolar assembly and function during interphase and also nucleolar disassembly during mitosis. The carboxyl-terminal two-thirds contains four consensus RNA-binding domains (CS-RBD), each about 80 amino acid residues long. The fourth CS-RBD is followed by a GAR (or RGG) domain of about 100 residues very near the carboxyl terminus. As in several other nuclear RNA-binding proteins (see previous discussion), the arginine residues within this GAR domain are dimethylated. Despite intense investigations of nucleolin, its precise nucleolar associations and functions remain uncertain. Nucleolin has been called a ribosome assembly factor (29). By sucrose gradient centrifugation, Herrera and Olson (33) showed that nucleolin associates with nascent rRNA, and it is believed that nucleolin binds nascent pre-rRNA in vivo. Recent experiments (34) showed that nucleolin decorates the RNP fibrils of Miller-spread CHO nucleolar chromatin, and it was shown by SELEX amplification that nucleolin preferentially binds a stem-loop structure consisting of a 4- to 5-bp stem and an eight-nucleotide loop containing a specific UCCCGA sequence motif. Interestingly, the six-nucleotide motif recovered by SELEX is also present in pre-rRNA, presumably at sites of nucleolin interaction (34). Binding to the motif is thought to be by one or more of nucleolin’s CS-RBDs. Nucleolin’s GAR domain may also function in RNA binding. It binds single-stranded DNA and RNA in vitro, and it displays RNA helicase (unwinding) activity.
Homologues of vertebrate nucleolin have been found in yeast. NSR1 in S. cerevisiae is an abbreviated version of the vertebrate protein. It also contains alternating acidic and basic domains at its amino terminus, but only two CS-RBD, followed by a GAR domain very near the carboxyl terminus. Gar2+ is the nucleolin homologue in S. pombe . Although the function of vertebrate nucleolin remains obscure, genetic knockout of these yeast homologues (35-37) disrupts processing of 18 S rRNA, causes a relative decrease in the abundance of 40 S small ribosomal subunits, and slows cell growth. These latter findings indicate that NSR1 and gar2+ have at least an associative role in pre-rRNA processing.
1.3.3. Nopp140, SRP40
Rat Nopp140 (SRP40 in yeast) is another nucleolar-specific phosphoprotein that contains alternating acidic and basic domains (38). Its acidic domains are extremely rich in serine residues that are also phosphorylated by casein kinase II. Its basic regions are similar to those of nucleolin, and it has substantial numbers of alanine, valine, lysine, and proline residues. Several potential MPF sites that have the amino acid sequence – /sPKK- are apparent in these basic domains. Similar to nucleolin, the highest steady-state concentrations of Nopp140 reside within the DFR, and to a lesser extent within the GR (39). One of Nopp140′s distinguishing characteristics, however, is its ability to shuttle from nucleoli to the cytoplasm along nuclear tracks that begin within the DFR (39). Nopp140 was first identified as a nuclear localization signal (NLS)-binding protein (40). The supposition is that Nopp140 acts as a nucleolar chaperone by shuttling NLS-containing proteins into the nucleolus. Nopp140 specifically binds wild-type NLS sequences (not mutated versions) but only when the acidic regions are fully phosphorylated. The dephosphorylated version of Nopp140 does not bind wild-type NLS.
1.3.4. NO38, B23, Nucleophosmin, and Numatrin
Similar to nucleolin, vertebrate NO38 (M of 38,000; also called B23, nucleophosmin, and numatrin) is a putative ribosomal assembly factor (29) and a nucleolar shuttling protein (31). Early reports showed that NO38 is enriched in the GR, but now it appears that NO38 localizes to both the DFR and GR (41). The function of NO38 also remains poorly understood. NO38 preferably binds single-stranded versus double-stranded nucleic acids. It binds RNA just as readily as it binds DNA. It binds nucleic acids cooperatively, and it destabilizes RNA helices (42). Interestingly, two isoforms of NO38 exist in rat (called B23.1 and B23.2), the isoforms result from alternative splicing of mRNA transcribed from one gene copy, and the B23.1 protein contains an additional 5-kDa tail at its carboxyl terminus that is not present in B23.2 (43). B23.1 is the predominant form that binds nucleic acids, and it resides within nucleoli. Conversely, B23.2 binds nucleic acids only weakly, and it resides within both the cytoplasm and the nucleoplasm but not within nucleoli. The functional significance of the two isoforms remains unknown but intriguing. Also like nucleolin, NO38 contains -Thr-Pro-Ala-Lys- (-TPAK-) sequences that are phosphorylated by MPF (44), thus suggesting that mitosis-specific phosphorylation of NO38 may help disassemble nucleoli (see later).
1.3.5. Other proteins
Several other nucleolar proteins have been identified, mostly in yeast (45). The yeast Ssb1 protein contains an RNA-binding CS-RBD, and a GAR domain, followed by an acidic region. It copurifies with snR10 and snR11 small nucleolar RNA, but its role in pre-rRNA processing remains unknown. Gar1 in S. cerevisiae is closely associated with 18 S rRNA processing and, as its name implies, also contains a glycine/arginine-rich domain. Nop2p in S. cerevisiae (46) is very similar to vertebrate p120, which is closely associated with cell proliferation (47). Nop3p is another GAR-containing nucleolar protein in S. cerevisiae. When it is genetically depleted, yeast cells show abnormal processing of pre-rRNA, when the 27SB intermediate (precursor of the 25S mature rRNA) and the 23S intermediate (precursor of the 18S mature rRNA) both accumulate. This results in a drop in producing both small and large ribosomal subunits and in slow growth (48). Nop4p in S. cerevisiae is required for formation of the 60 S large ribosomal subunit (49). Nop77p is a yeast nucleolar protein that mediates interactions between Nop1 (fibrillarin) and pre-rRNA (50). Sof1p is yet another protein associated with Nop1 and U3 in 18 S rRNA processing. Several nucleolar RNA-dependent helicases in yeast (Spb4p, Drs1p, Dbp3p, and CA9) are necessary for 40S and 60S ribosomal subunit assembly (49). Pop1 and Snm1 associate with the 7-2/MRP RNA in 5.8S processing (see later). The interplay in processing between all of these proteins with the pre-rRNA and with the small nucleolar RNA (see later) remains imperfectly understood.
1.4. Small Nucleolar RNA and Pre-rRNA Processing
The nascent pre-rRNA transcript (47S in mammals) is processed by a series of exo- and endonucleolytic cleavage reactions to yield intermediates of discrete size before the formation of mature 18 S, 5.8 S, and 28 S ribosomal RNA (51). The transcript is modified posttranscriptionally by methylation of bases at conserved sites and by the conversion of specific uridine residues to pseudo-uridine. The assembly with in-coming ribosomal proteins to form the ribosomal subunits is concurrent with cleavage and base modification. Several small nucleolar RNAs (snoRNA) have been identified that play either direct or associative roles in the processing reactions (45, 52). SnoRNAs in turn interact with nonribosomal nucleolar proteins (many described above) to form RNP complexes that engage in cleavage or base modification.
Initial cleavage of the mammalian 47S pre-rRNA at the 3′-T1 site yields the abundant 45S intermediate. The best characterized cleavage event, however, occurs within the 5′-ETS region of the primary transcript (53-55). U3 (56) and fibrillarin associate closely within the RNP complex known as "terminal balls" (54) that is positioned on the 5′-ETS (57). This terminal complex makes the first site-specific endonucleolytic cleavage in the processing of the 18S RNA (58). Miller spreads of nucleolar transcription units clearly show a relatively large RNP complex situated at the 5′-end of even the shortest fibrils. This RNP complex assembles on the 5′-ETS sequence specifically in the vicinity of cleavage sites at nucleotide 412 in human, 649 in mouse, or 782 in rat (57). Besides its association with the 5′-ETS cleavage, U3 is also associated with ITS1 and ITS2 cleavage reactions. Processing of 18S is inhibited by depletion of U3 (59), U14 and snR10 in yeast (60, 61), or of U14 and U22 in Xenopus. The 7-2/MRP snoRNA in yeast, with its associated Pop1 protein, plays a role in cleavage within ITS1 to liberate the yeast 5.8S rRNA. The U8 snoRNA in Xenopus mediates cleavage events within ITS2 and the 3′-ETS for proper excision of the 5.8S and 28S regions (62, 63).
SnoRNAs display conserved sequence elements that determine their conserved three-dimensional structures (45). For instance, the highly characterized U3 snoRNA contains conserved motifs designated boxes A, B, C, C’, and D. Boxes A, B, and C are unique to U3. Boxes A and B interact with rRNA, and box C associates with fibrillarin. Boxes C’ and D of U3 are common to many other snoRNAs that constitute a class called C/D snoRNA (box C’ of U3 is simply called C in these other snoRNAs). This box C (C’ in U3) is a conserved motif of six nucleotides that is six nucleotides downstream of the 5′-end of the snoRNA. Box D, on the other hand, is a conserved motif of four nucleotides that is just upstream of the 3′-end of the snoRNA. The 5′- and 3′-ends of the snoRNA form a short terminal double helix. Complementarity does not extend into boxes C and D, so they remain single-stranded and positioned next to the terminal helix. In the case of U14, boxes C and D are necessary for proper intron processing, stability, and accumulation (64). Interestingly, the majority of C/D snoRNAs are thought to associate with fibrillarin, again via box C.
More than fifty snoRNAs have been identified (65). Whereas snoRNA U3, U8, and U13 are encoded by their own independent genes, many of these other snoRNAs are encoded as introns of several different pre-mRNA molecules. For example, mouse U14 was the first snoRNA identified as part of an intron within the gene encoding the 70-kDa cognate heat shock protein hsc70 (66). Interestingly, many snoRNAs are transcribed as introns of pre-mRNA that encode proteins necessary for ribosomal assembly or factors necessary for protein translation (45). For example, snoRNA U17 in Xenopus is repeated in each of six introns of the pre-mRNA that encodes ribosomal protein S8 (67), and snoRNA U24 in humans and chickens is contained within an intron of the ribosomal protein L7 pre-mRNA. Both U20 and U23 in humans, mice, rats, hamsters, frogs, and fish are contained within introns of the nucleolin pre-mRNA. Expression of intronic snoRNA within these various pre-mRNA suggests a potential mechanism to coordinate synthesis of pre-rRNA processing components, ribosome structural proteins, and even translation factors. The intron processing mechanism that removes the snoRNA from the intron remains obscure. In the case of U14, the snoRNA region of the intron forms a loop whose ends are held together by a short double-helical stem of about seven base pairs. This helix actually spans what will be the final 5′-and 3′-ends of the snoRNA. An endonucleolytic cut within the stem is necessary to liberate the snoRNA from the rest of the intron. Characteristic of the intronic snoRNA, the cut is staggered, leaving a 3′-overhang of one base.
Although snoRNA U3, U8, U14, and U22 associate with fibrillarin in various pre-rRNA cleavage events, the processing roles for many of these other C/D snoRNAs have only recently been described (65, 68), and these results provide exciting discoveries about posttranscriptional methylation of the pre-rRNA. Just upstream of the D box in a particular snoRNA is a single-stranded region that is complementary to a specific pre-rRNA segment, either within the 18S, 5.8S, or 28S regions, but not to any of the spacer regions. The D box, together with an upstream helix formed between the snoRNA and the complementary segment of the pre-rRNA, directs the methylation specifically of the fifth nucleotide upstream of the D box. Methylation occurs on the ribose ring as 2′-O-methylation. Approximately 105 conserved methylation sites are known within the rRNA. Therefore, the majority of the C/D snoRNAs are guide RNAs, perhaps in association with fibrillarin, for site-specific methylation of the mature ribosomal RNA. Methylation creates hydrophobic sites within the rRNA that may direct proper RNA folding, processing, and subsequent interaction with incoming ribosomal proteins as ribosomal subunits assemble.
1.5. Nucleolar Cycle
The cycle of nucleolar assembly and disassembly is intimately linked to the cell cycle. The nucleolus disassembles during the prometaphase of mitosis and begins to reassemble during the telophase. The onset of mitosis is controlled by tight regulation of MPF, or cyclin B/p34cdc2 kinase. Once activated, MPF phosphorylates many different cellular substrates. At least two nucleolar proteins, nucleolin and NO38, are direct substrates for MPF (44). The phosphorylation sites in hamster nucleolin are the threonine residues within nine -TP A/qKK- motifs found within the basic domains of the amino terminus. At least one other nucleolar protein of known sequence, Nopp140, displays potential MPF phosphorylation sites. The current supposition is that MPF-specific phosphorylation of key structural proteins within the nucleolus contributes to nucleolar disassembly.
Another prominent factor in nucleolar disassembly may be the shutdown of rDNA transcription. Interphase nucleoli partially disassemble when actinomycin D blocks rDNA transcription (23). Transcription of the pre-rRNA genes normally begins to decline in the prophase. This decline may result from tight compaction of rDNA into the nucleolar organizer regions of the mitotic chromosomes. Immunofluorescence studies of mitotic cells show that at least fibrillarin and nucleolin, two proteins involved with pre-rRNA processing and packaging, diffuse into the cytoplasm (the nuclear envelope has disassembled by this time) upon nucleolar disassembly. SnoRNA U8 also diffuses to the cytoplasm (69). The lack of pre-rRNA transcripts may contribute to the diffusion of RNA processing components. Interestingly, RNA polymerase I and the transcription factors UBF and SL1 all remain bound to the nucleolar organizer regions of the chromosomes (70).
Nucleolar reassembly (nucleologenesis) begins during the telophase with the appearance of prenucleolar bodies. These particles contain fibrillarin, NO38, nucleolin, Nopp140, and the snoRNA U3, but no rRNA. All are factors closely associated with the processing of pre-rRNA. As rDNA transcription by polymerase I resumes in the telophase, these prenucleolar bodies coalesce with the nucleolar organizers to reconstitute intact nucleoli by the interphase. By the early interphase, there may be several functional nucleoli, which may fuse to form fewer but larger nucleoli. Interestingly, the prenucleolar bodies share morphological and compositional similarities with coiled bodies (71), originally described in 1903 as accessory bodies of nucleoli in nerve cells (72). Coiled bodies contain fibrillarin, ribosomal protein S6, DNA topoisomerase I, and snoRNA U3, but no rRNA. In addition, coiled bodies contain the specific p80 coilin protein and snRNPs normally involved with pre-mRNA splicing. The precise relationship between prenucleolar bodies and coiled bodies remains imperfectly understood.
