Regulation of ribosome synthesis is an essential aspect of growth control. Thus far, little is known about the factors that control and coordinate these processes. We show here that the Caenorhabditis elegans gene ncl-1 encodes a zinc finger protein and may be a repressor of RNA polymerase I and III transcription and an inhibitor of cell growth. Loss of function mutations in ncl-1, previously shown to result in enlarged nucleoli, result in increased rates of rRNA and 5S RNA transcription and enlarged cells. Furthermore, ncl-1 adult worms are larger, have more protein, and have twice as much rRNA as wild-type worms. Localization studies show that the level of NCL-1 protein is independently regulated in different cells of the embryo. In wild-type embryos, cells with the largest nucleoli have the lowest level of NCL-1 protein. Based on these results we propose that ncl-1 is a repressor of ribosome synthesis and cell growth.
During the life cycle of metazoans, growth occurs by increases in both the size of individual cells and in the actual number of cells. Most often these two processes are coordinated; cells are born at a certain size, double in size during the cell cycle, and then divide (Prescott, 1976). In some cases, however, cell size increase and cell division occur independently, as in early embryogenesis of many organisms where rapid cleavage divisions occur without any increase in cell size (Wilson, 1896). Similarly, growth of individual cells can occur in the absence of cell division. For instance, many adult tissues such as cardiac muscle increase in mass by cell size increase, or hypertrophy, rather than by cell division (Rakusan, 1984). A striking example of growth without division is during amphibian oocyte development (Dumont, 1972). Furthermore, an increase in cell size independent of cell division may also account for the fact that distinct cell types having identical DNA content can vary greatly in size within an individual organism (Altman and Katz, 1976).
Previous studies in bacteria and yeast have demonstrated that cell growth (increase in cell size and number) is tightly coupled to an increase in both protein synthesis and ribosome biogenesis (for review see Nomura et al., 1984). In higher eukaryotes, this correlation is seen in the diminished transcription of ribosomal RNAs (rRNA and 5S RNA) during quiescence and the subsequent increase after stimulation with growth factors (Johnson et al., 1974). Cell division is not a requirement for this coordination, however, as studies of hypertrophy in vertebrate cardiomyocytes also link rRNA and 5S RNA synthesis with control of cell size (McDermott et al., 1989). These studies suggest that regulators of ribosome synthesis may play an important role in cellular growth control. One gene that appears to function in the control of cell growth is the Retinoblastoma gene (Rb).1 In addition to its role in the control of cell division, Rb has been shown to inhibit the synthesis of both rRNA and 5S RNA (Cavanaugh et al., 1995; White et al., 1996).
In eukaryotes, the nucleolus, the nuclear organelle in which rRNA synthesis and assembly of preribosomal subunits occurs (for review see Melese and Xue, 1995), is a good cytological marker of ribosome synthesis as the size of the nucleolus is indicative of the level of rRNA synthesis (Kurata et al., 1978; Altmann and Leblond, 1982; Moss and Stefanovsky, 1995). Interestingly, tumor cells not only exhibit rapid cell growth and division, but also often have enlarged nucleoli (Busch and Smetana, 1970; Derenzini and Trere, 1991). Similarly, during both normal developmental cardiac hypertrophy and aberrant hypertrophy that occurs in many forms of heart disease, nucleoli are enlarged (Hatt et al., 1978; Cluzeaud et al., 1984; Dalen et al., 1987). Enlarged nucleoli have also been observed in the ncl-1 mutant of the nematode Caenorhabditis elegans (Hedgecock and Herman, 1995). We have isolated the ncl-1 gene and show that it appears to repress rRNA and 5S RNA synthesis and to be an inhibitor of both ribosome biosynthesis and cell growth.
Materials And Methods
Strains and Alleles
Bristol strain N2 was used as the standard wild-type strain. The ncl-1 alleles used were: e1865 (Hedgecock and Herman, 1995) and e1942 (provided by C. Kenyon, University of California, San Francisco, CA). Both were outcrossed against N2 five times. clr-1 (e1745) was used for cell size analysis. Nematode strains were cultured as described by Brenner (1974).
Worm Size and Total Protein Measurements
Two-dimensional images of wild-type and ncl-1 (e1942) worms grown for 50 h at 20°C after hatching were collected using a charge-coupled device camera and National Institutes of Health (NIH) image. Worm sizes were determined in Canvas (Deneba Software, Miami, FL) and then analyzed in Microsoft Excel (Redmond, WA). The amount of total protein was determined for wild-type and ncl-1 (e1942) adult worms grown at 20°C for 44 h after hatching using the Amidoschwarz assay (Schaffner and Weissmann, 1973). Total DNA in each sample was determined using a fluorometer (model TKO-100; Hoefer Scientific Instruments, San Francisco, CA). Analysis of total protein to DNA ratios and statistics was performed using Microsoft Excel.
In Situ Hybridization
In situ hybridization using either 5.8S or internal transcribed spacer (ITS)2 probes was done according to the method of Seydoux and Fire (1994). The 5.8S (GGAGGCCCAGTTGGTGCTATGCGTTCG) and ITS2 (CAAGTCGGACGCCATTTGGCAGCTCCTCCG) antisense oligo probes were directed against nucleotides 2204–2230 and 2426–2455, respectively, of the C. elegans rDNA repeat (Ellis et al., 1986). Probes were end labeled with terminal deoxynucleotidyl transferase and digoxigenin– ddUTP (Boehringer Mannheim Biochemicals, Indianapolis, IN) and then used at a concentration of 1.0 μg/ml in 7.7% formamide, 2× SSC hybridization buffer. Probes were then visualized using FITC-conjugated antidigoxigenin antibodies (Boehringer Mannheim Biochemicals). The following controls for specificity of hybridization were performed: an anti-SL1 (spliced leader) probe (CTCAAACTTGGGTAATTAAACC) hybridized to the cytoplasm of all cells and to cytoplasmic granules in the P cells as previously described (Seydoux and Fire, 1994), but did not hybridize to nucleoli; and prehybridizing of slides with 20 μg/ml of unlabeled 5.8S or ITS2 antisense oligo inhibited hybridization with the respective labeled oligos.
Run-on Transcription Assays
Extracts were prepared from early embryos (95% earlier than 100-cell stage) by collecting and bleaching worms from the different genetic backgrounds just as they began to become gravid and contained <10 embryos per hermaphrodite. Preparation of extracts for transcription assays and quantitation of DNA in the extracts was done as described (Schauer and Wood, 1990), except that salmon sperm DNA was used to provide a standard curve and a fluorometer (model TKO-100; Hoefer Scientific Instruments) was used.
For some experiments, run-on transcription reactions and hybridizations were performed as described (Schauer and Wood, 1990), except that volumes of embryonic extracts corresponding to 50 μg of DNA, 100 μCi of [32P]uridine 5′-triphosphate (UTP), 2 mM ATP, 2 mM CTP, 2 mM GTP, and 0.05 mM cold UTP were used in 100-μl reactions and then hybridizations were done for 2 d at 42°C. In other experiments, the reaction conditions were as follows: volumes of extract corresponding to 2.5 μg of DNA were incubated in buffer containing 5 mM Tris-HCl, pH 8.0, 2.5 mM MgCl2, 150 mM KCl, 100 μCi [32P] cytosine 5′-triphosphate (CTP), 0.8 mM ATP, 0.8 mM GTP, 0.8 mM UTP, 1 mM DTT, and 200 U/ml RNAsin (Promega Corp., Madison, WI) (final volume was 20 μl) for 30 min at room temperature. For these experiments, hybridization was performed for 18 h at 65°C in 10 mM TES (Sigma Chemical Co., St. Louis, MO) pH 7.5, 1% SDS, 10 mM EDTA, 250 μg/ml Escherichia coli RNA, 300 mM NaCl, 100× Denhardt's (2% Ficoll 400, 2% polyvinylpyrrolidone, 2% bovine serum albumin), and 0.25% powdered nonfat milk. In each experiment, three reactions were performed for each genotype tested.
Slot blot filters were prepared as follows: for each reaction performed, 5 μg of each of the DNAs of interest were denatured in 0.3 M NaOH at 65°C for 10 min and then neutralized by the addition of 1 vol of 2 M ammonium acetate. Denatured DNAs were loaded onto Gene Screen Plus membrane (DuPont-NEN, Boston, MA), presoaked in 1× SSC, 10 mM Tris, pH 7.4, using a slot blotting apparatus (Schleicher & Schuell, Inc., Keene, NH). DNAs of interest were as follows: rDNA was pCe7 (Files and Hirsh, 1981); 5S RNA gene construct was made by cloning a 160-bp StuI-HindIII fragment of plasmid 985(Sl1/5S) (Ferguson et al., 1996) into Bluescript KS+ (Stratagene, La Jolla, CA) cut with EcoRV and HindIII; histone genes were in clone pCeh-3 (Roberts et al., 1989). Quantitation of hybridization signals was performed using a phosphorimager (Molecular Dynamics, Sunnyvale, CA) and data from all experiments performed (four experiments for wild-type and ncl-1 [e1865] on rDNA and 5S; two experiments for ncl-1 [e1942] on rDNA, 5S, and histones, and wild-type and ncl-1 [e1865] on histones) were collated for the graphs presented in Fig. 3.
Northern Blot Analysis
Northern blots were performed as described in Ausubel et al. (1987). The 28S rRNA (CTGCACTGACGGAAGCTCCAGCCGAGCC) antisense oligo probe was directed against nucleotides 4113–4140 of the C. elegans rDNA repeat (Ellis et al., 1986). The histone (AGAGGGCCGTTGGGTTCGGT) antisense oligo probe was directed against the conserved 3′ sequence in all C. elegans histone messages (Roberts et al., 1989).
Cloning of ncl-1
The ncl-1 locus was identified through a series of germline transformation experiments with cosmids and subclones of cosmid ZK112. DNA was injected with the dominant marker rol-6 (Mello et al., 1991) into the syncytial gonad of ncl-1 (e1865) or ncl-1 (e1942) animals. DNA was injected at a final concentration of 0.1 mg/ml. Adult Rol progeny were analyzed by differential interference contrast microscopy for rescue of the Ncl phenotype. Partial rescue (not all cells were rescued) was obtained with the 7.5-kb rescuing fragment (pZK112-1b).
Antisense RNA derived from a 1-kb HindIII-XhoI fragment of pZK112-1b was injected into the syncytial gonad of wild-type hermaphrodites according to the procedure of Guo and Kemphues (1995). These worms gave progeny with enlarged nucleoli, indicative of ncl-1 phenocopy.
A full-length ncl-1 cDNA was obtained by screening a C. elegans mixed stage library (Stratagene) with the 1-kb HindIII-XhoI fragment of pZK112-1b. Sequencing of this cDNA revealed the same exon boundaries as predicted by GeneFinder. This cDNA begins at a position 134-bp upstream of the first ATG and terminates with a poly A tail 1,355-bp downstream of the first in-frame stop codon. The Matcher program (Fischetti et al., 1993) was used to identify the predicted coiled-coil motif in NCL-1. Sequence data is available under EMBL/GenBank/DDBJ accession number AF047027.
Production of Monoclonal Antibody, Immunoblotting, and Immunostaining
The pET-16b vector (Novagen Inc., Madison, WI) was used to generate a protein fusion between the 10-His tag and amino acids 24–377 of NCL-1. The bacterially expressed fusion protein 10-His–NCL was purified by chromatography on Ni2+-NTA-Agarose (QIAGEN Inc., Santa Clarita, CA). mAb D3C2 and ascites fluid were generated as described by Harlow and Lane (1988).
For immunoblotting, embryo extracts prepared by overnight extraction of embryos in N,N-dimethylformamide (Sigma Chemical Co.) followed by sonication of insoluble material in NEST-2 (5% SDS, 20 mM EDTA, 50 mM Tris-HCl, pH 6.8) were run on 6% SDS–polyacrylamide gels and then transferred to nitrocellulose. Blots were stained with Ponceau S, blocked for 1 h in 3% BSA and 10% nonfat dry milk in TBST (50 mM Tris-Hcl, pH 8.0, 150 mM NaCl, 0.05% Tween 20), incubated overnight at room temperature in primary antibody, washed in TBST, incubated in horseradish peroxidase-linked anti–mouse Ig (Amersham Corp., Arlington Heights, IL) diluted 1:2,500 in 3% BSA in TBST, washed extensively in TBST, and then processed for detection using the SuperSignal Substrate, Western blotting kit (Pierce Chemical Co., Rockford, IL). mAb D3C2 was used as undiluted hybridoma supernatant.
Fixation of embryos was as described (Albertson, 1984) with the following exceptions. Slides were pretreated with 3-aminopropyltriethoxysilane (Sigma Chemical Co.) and fixation was then done in N,N-dimethylformamide at −20°C for 5 min. Gonad fixation and staining was performed in solution. For immunostaining, mAb D3C2 ascites fluid was used at a 1:500 dilution. mAb K121 against 2,2,7-trimethylguanosine (Oncogene Science, Inc., Manhasset, NY) was used at a 1:50 dilution.
Increased Nucleolar and Cellular Size in ncl-1 Mutants
To begin to investigate the mechanisms regulating cell growth and ribosome synthesis, we have focused on the gene ncl-1 in C. elegans. ncl-1 (e1865) and ncl-1 (e1942) have previously been described as recessive mutations that result in enlarged nucleoli that can be detected by Nomarski microscopy in nearly all cells of live worms (Hedgecock and Herman, 1995). Both alleles are recessive and appear genetically to be loss of function mutations. This phenotype is fully penetrant in both alleles and is especially evident in embryos. For example, nucleoli are clearly detectable in a ncl-1 four-cell embryo but are not visible in a wild-type four-cell embryo (Fig. 1). In larvae and adult worms, enlarged nucleoli can be seen in all cells with the exception of the gut and syncytial gonad since these tissues have large nucleoli in wild type. Similarly, the degree to which nucleolar size changes in the mutant differs in distinct cell types; the small or not detectable nucleoli of some wild-type cells are clearly enlarged in ncl-1 mutants, whereas the large nucleoli of other cells are only slightly enlarged in ncl-1 mutants.
In addition to enlarged nucleoli, ncl-1 mutants also exhibit increases in growth. By microscopic imaging and measurement of worms in two dimensions, we observed that adult ncl-1 (e1942) worms are 9% larger than wild-type worms at identical times after hatching (n = 20 for each, P < 0.01). Previous studies have shown that there is no difference in cell division timing or final number of cells between ncl-1 mutants and wild type (Hedgecock and Herman, 1995). The enlarged size of ncl-1 mutants, therefore, is likely due to increased cell sizes. To test this, we used Nomarski microscopic imaging of live worms to analyze the sizes of specific cells and found that some cells in ncl-1 mutants are larger than the respective cells in wild-type worms. To quantitate the size increase, we made use of the mutant clr-1; these mutant worms are clear when shifted to the nonpermissive temperature, thereby allowing cell outlines to be visualized and cell sizes to be measured (Kipreos et al., 1996). We made two-dimensional measurements of specific cells and compared their sizes between clr-1 and ncl-1 (e1942);clr-1 larvae at identical times after hatching. The cell anterior lateral microtubule (ALM) is 47–78% larger in ncl-1;clr-1 than in clr-1, and the cell Q2.pap is 75–126% larger (Table I). The evidence that the ncl-1 gene product functions in these two cells is that they both have enlarged nucleoli in ncl-1 and ncl-1;clr-1 worms (data not shown), and that ncl-1 is known to function cell autonomously (Hedgecock and Herman, 1995). These results demonstrate that ncl-1 mutants have larger cells than wild-type worms and suggest that the ncl-1 gene product represses cell growth.
Increased Protein and rRNA in ncl-1 Mutants
Because an increase in cell size is likely to indicate increased protein synthesis, we assayed protein levels in wild-type and mutant worms and found that adult ncl-1 (e1942) worms have 22% more protein than wild-type worms (n = 10 for each, P < 0.0005). Because this increased level of protein in ncl-1 worms may indicate an increased capacity for protein synthesis, or more ribosomes, and because ncl-1 mutants have enlarged nucleoli, we hypothesized that ncl-1 mutant worms might have an increased amount of ribosomal RNA. To quantitate steady-state rRNA, we performed Northern blots on RNA from mutant and wild-type worms 48 h after hatching. We found that the ratio of histone mRNA to total DNA is the same in ncl-1 and wild-type worms at identical times after hatching, and therefore standardized the 28S rRNA hybridization signal to the histone message on the same blot. We found that ncl-1 (e1865) adults contain 1.6 times more rRNA than wild-type worms (P < 0.05) and that ncl-1 (e1942) adults contain 2.2 times more rRNA than wild-type worms (P < 0.05) (Fig. 2). Thus, one way in which the ncl-1 gene product could restrict cell size is by repression of the steady-state level of rRNA and, therefore, the capacity for protein synthesis.
Increased RNA Polymerase I and III Transcription in ncl-1 Mutants
We next considered the possibility that the increased amount of rRNA in ncl-1 mutants is due to an increase in the rate of rRNA transcription. To test this hypothesis, we compared the level of rRNA present in the nuclei of early wild-type and ncl-1 embryos by in situ hybridization. Although we detected 5.8S rRNA in the cytoplasm in both wild-type and ncl-1 embryos, hybridization in the nucleus was higher in ncl-1 embryos than in wild-type (Fig. 3, a and b). In all ncl-1 embryos examined, nuclear hybridization was seen in one or two sites that overlapped with the sites of visible nucleoli, whereas this pattern was not seen in any of the wild-type embryos examined (Fig. 3,b). To distinguish between precursor and mature rRNA, we next performed in situ hybridizations using an oligo probe directed against ITS2 that is present in precursor rRNA but is removed during processing (Perry, 1976). No signal was detected in the cytoplasm of either wild-type or ncl-1 embryos using this probe. In the nucleus, hybridization was much stronger in ncl-1 than in wild-type embryos (Fig. 3, e and f). Similar to the 5.8S probe, we detected one or two sites of hybridization that colocalize with nucleoli in nuclei of all ncl-1 embryos examined (Fig. 3 f) This was not seen in any of the wild-type embryos examined. These data indicate that in ncl-1 embryos, there is an increase in the level of precursor rRNA present in the nucleus.
To determine if this increase in level of precursor rRNA in the nucleus is due to an increase in the rate of rRNA transcription in mutant embryos, we performed nuclear run-on assays on wild-type and mutant embryos (refer to Materials and Methods). We analyzed early embryos because at this time in development there is a very clear difference between the size of nucleoli in mutant and wild-type (refer to Fig. 1). Nuclear run-on assays were performed on equivalent numbers of nuclei from mutant and wild-type extracts as determined by standardization to DNA by fluorometry. We first analyzed rRNA transcription by hybridization of nascent RNA labeled in the run-on reaction to DNA corresponding to one repeat of C. elegans rDNA. We found that the rate of rRNA transcription is 2–2.5 times higher in ncl-1 early embryos than in wild-type embryos (Fig. 4 a). This observation suggests that, as hypothesized by Hedgecock and Herman (1995), the ncl-1 gene product represses rRNA transcription, and that ncl-1 worms have enlarged nucleoli because of increased nucleolar activity.
To address the polymerase specificity of the ncl-1 phenotype, we assayed RNA polymerase II and III transcription using the nuclear run-on assay. To detect RNA polymerase III transcription, we hybridized radiolabeled nascent RNA to DNA encoding the C. elegans 5S gene. RNA polymerase II transcription was detected by hybridization of radiolabeled nascent RNA to DNA encoding the histone H2A, H2B, H3, and H4 genes. We found that 5S transcription was approximately twofold higher in ncl-1 than in wild-type embryos (Fig. 4,b), whereas histone transcription was not significantly affected (Fig. 4 c). These data indicate that the ncl-1 gene product functions to modulate both RNA polymerase I (rRNA) and RNA polymerase III (5S) transcription, whereas it does not appear to play a general role in the regulation of RNA polymerase II transcription. This suggests that ncl-1 may repress the production of ribosomes, and thus the capacity for protein synthesis, through inhibition of RNA polymerase I and III transcription.
NCL-1 Is a B Box Zinc Finger Protein
ncl-1 had previously been localized to cosmid C33C3 by transformation rescue (Miller et al., 1996). We achieved rescue with the neighboring cosmid ZK112, which has been sequenced by the C. elegans genome consortium (Sulston et al., 1992). Analysis of subclones from this cosmid resulted in identification of a 7.5-kb genomic fragment containing one predicted gene that partially rescued the ncl-1 enlarged nucleolus phenotype (Fig. 5). To confirm that this gene is ncl-1, we injected antisense RNA transcribed from this genomic fragment into the syncytial gonad of wild-type hermaphrodites. This resulted in enlarged nucleoli in larval progeny (Fig. 5). In both the rescue and antisense experiments, the numbers of affected nuclei varied from animal to animal. These lines of evidence together indicate that this gene is ncl-1.
The ncl-1 cDNA can encode an 851-amino acid protein (Fig. 6 a) with two zinc finger motifs previously described as B boxes (Reddy et al., 1992). It also contains a predicted coiled-coil motif. Proteins that share these motifs include the oncogenic proteins promyelocytic leukemia (PML), T18, and ret finger protein (RFP), as well as other proteins such as XNF7, SS-A/Ro, regulatory protein, T lymphocyte (RPT-1), and PwA33 (for review see Reddy et al., 1992). Although NCL-1 does not contain a really interesting new gene (RING) zinc finger motif common to many of these proteins, it is most like PML and T18 in that these proteins all have two B box zinc finger motifs. Outside of the B boxes and coiled-coil motif, there is no homology between NCL-1 and any other proteins.
Expression of NCL-1 Is Correlated with Repression of rRNA Synthesis and Decreased Cell Size in Wild-type Embryos
We were interested in learning the spatial and temporal expression of NCL-1 and therefore generated mouse polyclonal antisera and a mouse mAb against the amino-terminal half of NCL-1 (refer to Materials and Methods). Results with polyclonal antisera and the monoclonal antibody are identical; only those experiments using mAb D3C2 are presented here. mAb D3C2 recognizes two bands of ∼100-kD in wild-type extracts, but not in extracts made from either of the two mutant alleles of ncl-1 (Fig. 6 b). The upper band is a phosphorylated form of NCL-1 (data not shown). The predicted molecular weight of NCL-1 is 92-kD. In ncl-1(e1865) embryo extracts, a faster migrating polypeptide of ∼80-kD was detected, whereas no immunoreactive species was detected in extracts from ncl-1 (e1942). These results confirm that this gene is ncl-1 and suggest that ncl-1(e1942) is a protein-null allele, whereas ncl-1(e1865) produces a shorter form of the protein.
Given that NCL-1 appears to act as an inhibitor of RNA polymerase I transcription, and thus nucleolar size and function, we hypothesized that it would be expressed when nucleoli are absent, and not expressed when nucleoli are visible. We first examined early embryos in which nucleoli are not detectable in wild-type. Indirect immunofluorescence using mAb D3C2 showed that in wild-type embryos, NCL-1 is a cytoplasmic protein with no obvious nuclear accumulation. NCL-1 staining is abundant in early embryos and then rapidly diminishes during development (Fig. 7 a). We detect no staining with mAb D3C2 in ncl-1 (e1942) mutant embryos at any stage (data not shown).
We next assayed NCL-1 expression in 300-cell stage embryos, since nucleoli are first visible in some cells at this stage in development. As a marker for nucleoli in these cells, we stained similarly staged embryos with mAb K121 that recognizes nucleoli in C. elegans (Fig. 7,d). Immunostaining with mAb D3C2 showed that NCL-1 is present in the cytoplasm of all cells except those of the developing gut (Fig. 7,c); these cells are among the first to differentiate in the embryo, are the largest cells in the embryo, and have very large nucleoli. Lastly, we analyzed the NCL-1 staining pattern in the hermaphrodite gonad. Nucleoli are visible in the syncytial gonad of adult hermaphrodites but disappear as oocytes mature. We observed that cytoplasmic NCL-1 staining is absent in the mitotic region of the syncytial gonad, low in the meiotic region, and then increases throughout oocyte maturation such that the most intense staining is seen in the most mature oocyte (Fig. 7 e). Together these staining experiments indicate that NCL-1 protein is present in cells in which nucleoli are absent, and absent from large cells in which nucleoli are prominent. This suggests that the regulation of NCL-1 protein expression is important for the control of cell growth and ribosome synthesis.
NCL-1, a Key Molecule in the Regulation of Cell Growth and Ribosome Synthesis
We have shown that the C. elegans gene ncl-1 may inhibit ribosome synthesis and restrict cell growth. In ncl-1 loss of function mutants, individual cells are larger than wild-type and as a result, the worms themselves are larger. In addition, we observed a twofold increase in steady-state amount of rRNA in ncl-1 mutant worms relative to wild-type. As rRNA is very stable (for example, Liebhaber et al.  reported a rRNA half-life of ⩾700 h in primary human fibroblasts), this increase is likely due to the 2–2.5 fold increase in the rate of rRNA transcription we observed in ncl-1 mutant embryos relative to wild-type. We observed a similar effect on the rate of 5S RNA transcription. These results are consistent with those seen by others in systems in which rRNA or 5S RNA transcription are experimentally induced, such as by application of 12-o-tetradecanoylphorbol-13-acetate to cells (Allo et al., 1991; Garber et al., 1991; Vallett et al., 1993). Furthermore, since ribosomal RNA makes up 60–90% of cellular RNA, a twofold increase is a substantial alteration. This result suggests that ncl-1 worms may have more ribosomes than wild-type worms, consistent with our finding that ncl-1 worms contain more protein than wild-type worms. Therefore, the difference in cell size and whole animal size between ncl-1 and wild-type worms may be accounted for by an increased ability to synthesize proteins.
The regulation of expression of NCL-1 protein appears to be important for the control of cell growth and ribosome synthesis as its abundance is inversely related to the presence of visible nucleoli, the nuclear organelles in which rRNA synthesis occurs. NCL-1 is present in the early embryo where nucleoli are not visible and high levels of zygotic rRNA production are perhaps not necessary because of the maternal contribution of ribosomes. At the 300-cell stage, NCL-1 protein is present in all cells except those of the gut. Gut cells are the first cells to differentiate in the embryo, are the largest cells in the embryo (Wood, 1988), and have large nucleoli. In addition to being very large, the gut cells are quite active metabolically (the gut produces all of the digestive enzymes as well as the yolk for the gonad; White, 1988) and therefore presumably require high levels of ribosome synthesis. Tissues with high metabolic activity, such as the gut, produce high levels of ribosomes as demonstrated in studies of Drosophila paragonial glands (Schmidt et al., 1985). Such cells likely synthesize high levels of ribosomes not solely to increase their size, but rather to maintain a high level of biosynthetic activity. Therefore, it is possible that the function of ncl-1 can be thought of not simply as being a repressor of cell size, but more specifically as an inhibitor of the level of biosynthetic activity within a cell. Together, our results indicate that NCL-1 is a key molecule in the regulation of cell growth and ribosome synthesis.
Possible Mechanisms for ncl-1–mediated Regulation of Ribosome Synthesis and Cell Size
The mechanism by which ribosome synthesis is coupled to cell growth is unclear. It has been proposed that a feedback mechanism exists wherein the ratio of free ribosomes to polysomes is monitored. When this ratio becomes too high (i.e., there are many inactive ribosomes not engaged in translation), ribosomal RNA synthesis is downregulated, and when this ratio becomes too low (i.e., the majority of ribosomes are engaged) then ribosomal RNA synthesis is upregulated (for review see Nomura et al., 1984). In light of this, two models for the molecular function of ncl-1 can be considered. The first is that ncl-1 functions to inhibit translation. This would result in a decrease in polysomes and an increase in free ribosomes, which would lead to a decrease in rRNA transcription. The fact that ncl-1 worms have more protein, and presumably a higher level of protein synthesis, than wild-type worms is consistent with this model. The second model is that ncl-1 functions in sensing the ratio of free ribosomes to polysomes and directs the level of rRNA transcription. Finally, it is possible that although the majority of NCL-1 protein appears by immunocytochemistry to be cytoplasmic, some NCL-1 protein may be nuclear, and might act directly as a transcription factor to inhibit RNA polymerase I and III transcription.
The control of cell size is of clinical interest as many forms of human heart failure and kidney disease caused by diabetes are associated with enlarged cells (hypertrophy) in cardiac and renal tissues (for review see Hannan and Rothblum, 1995; Kurabayashi and Yazaki, 1996; Rabkin and Fervenza, 1996). In studies of cardiac hypertrophy, it has been found that ribosomal RNA synthesis increases (McDermott et al., 1989), consistent with the idea that increased cell size is correlated with an increased capacity for protein synthesis. Although the mechanism of this transcriptional upregulation has not been determined, it has been found that UBF (upstream binding factor, the RNA polymerase I transcriptional activator) becomes phosphorylated (Hershey et al., 1995), an event which is correlated with its activation (Voit et al., 1992). Likewise, ribosomal protein S6 becomes phosphorylated by S6 kinase (Takano et al., 1996); phosphorylation of S6 correlates with higher translation efficiency (Proud, 1996). It is possible that this results in an increase in rRNA transcription via the feedback mechanism discussed above. Thus far, there are no genetic systems available to study hypertrophy. As ncl-1 mutants exhibit phenotypes characteristic of hypertrophy (increased cell size, increased steady-state rRNA, and increased rate of rRNA transcription), we propose that ncl-1 provides a model genetic system in which to approach this problem of the regulation of hypertrophic growth.
NCL-1 Is Related in Sequence to Oncogenes Such as PML and May Share with Rb the Ability to Repress the Synthesis of Ribosomal RNAs
NCL-1 has B box zinc finger and coiled-coil motifs common to several proteins including PML, T18, and RFP (for review see Reddy et al., 1992). This similarity is intriguing in that these three genes were all identified as translocations that result in tumorigenesis. For instance, PML was found as a translocation resulting in a fusion protein between the zinc fingers and coiled-coil of PML and the retinoic acid receptor (RAR); the fusion protein PML–RAR is responsible for acute promyelocytic leukemia in humans (Borrow et al., 1990; de The et al., 1990; Kakizuka et al., 1991). Although the mechanism of transformation by the PML–RAR fusion protein is still unclear, PML alone may function normally as a growth suppressor since it has been shown to be able to inhibit transformation in cell culture (Mu et al., 1994). Because tumor cells often have enlarged nucleoli (Busch and Smetana, 1970; Derenzini and Trere, 1991), it is conceivable that these genes and ncl-1 might have overlapping functions.
One other cancer related gene, the tumor suppressor Rb, is structurally unrelated but does share some common functions with ncl-1. Recent evidence suggests that, like ncl-1 mutants, loss of Rb function may derepress transcription by RNA polymerase I and III (Cavanaugh et al., 1995; White et al., 1996). It is possible that one way in which tumor cells are able to achieve an increase in cell growth is by increasing their ability to synthesize proteins through mutation of a gene such as Rb that normally functions to repress RNA polymerase I and III transcription (for review see Nasmyth, 1996; White, 1997). Interestingly, it has been shown that chimeric mice containing Rb-deficient cells have enlarged cells in the liver and cerebellum (Williams et al., 1994), suggesting that Rb, like ncl-1, may be required to limit the size of cells. Furthermore, induction of cellular hypertrophy in cardiomyocytes leads to increased phosphorylation of Rb (Sadoshima et al., 1997). It will be interesting to learn if there is Rb in C. elegans, and if so, how it and ncl-1 might function together to achieve the goal of negatively regulating cell size and ribosomal RNA synthesis. Conversely, it will be interesting to learn if there is a ncl-1-like gene in mammalian cells and whether or not its function is abrogated in tumor cells.
We thank our colleagues at Fred Hutchinson Cancer Research Center: B. Edgar, L. Moore, M. Morrison, S. Parkhurst, J. Priess, J. Roberts, L.B. Roth, and J. Stark for critical reading of the manuscript and D. Waring for helpful advice.
Abbreviations used in this paper
D.J. Frank was supported by a graduate fellowship from the National Science Foundation. This work was supported by a National Institutes of Health grant (GM48435-01A2) to M.B. Roth.
Address all correspondence to Mark B. Roth, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave., N Mailstop A3-013, Seattle, WA 98109. Tel.: (206) 667-5602. Fax: (206) 667-6877. E-mail: firstname.lastname@example.org