Physical and Functional Interaction of the p14ARF Tumor Suppressor with Ribosomes*

Alterations in the p14ARF tumor suppressor are frequent in many human cancers and are associated with susceptibility to melanoma, pancreatic cancer, and nervous system tumors. In addition to its p53-regulatory functions, p14ARF has been shown to influence ribosome biogenesis and to regulate the endoribonuclease B23, but there remains considerable controversy about its nucleolar role. We sought to clarify the activities of p14ARF by studying its interaction with ribosomes. We show that p14ARF and B23 interact within the nucleolar 60 S preribosomal particle and that this interaction does not require rRNA. In contrast to previous reports, we found that expression of p14ARF does not significantly alter ribosome biogenesis but inhibits polysome formation and protein translation in vivo. These results suggest a ribosome-dependent p14ARF pathway that regulates cell growth and thus complements p53-dependent p14ARF functions.

The INK4a/ARF locus on chromosome 9 is frequently altered in human cancer, and inherited INK4a/ARF mutations are associated with melanoma susceptibility in 20 -40% of multiple case melanoma families (1). This complex sequence encodes the melanoma tumor suppressor proteins, p16 INK4a and p14 ARF from alternative spliced transcripts in different reading frames (2,3). Both proteins are centrally involved in the regulation of cell cycle and apoptotic programs in response to oncogenic stimuli and are therefore frequently targeted in tumor development and progression (4). INK4a/ARF-deficient mice are tumor-prone (5)(6)(7), and those with melanocyte-specific expression of activated H-ras develop cutaneous melanomas with high penetrance (8). High density mapping on chromosome 9p in human melanomas has identified p14 ARF as the most commonly deleted INK4a/ARF gene (9), and individuals with altered ARF, but apparently wild type p16 INK4a , are melanoma-prone (10 -13).
Nucleolar ARF, rather than residing in inactive "storage" (20,21,28), may regulate ribosome biogenesis by retarding the processing of early 47 S/45 S and 32 S rRNA precursors (29). These effects do not depend on hdm2 or p53 but may involve the interaction of ARF with nucleophosmin/B23 in complexes of very high molecular mass (30,31). B23 is an abundant nucleolar endoribonuclease that is required for the nucleolar targeting of ARF (21) and for the maturation of 28S rRNA (32). Consistent with its role in inhibiting rRNA processing, ARF promotes the ubiquitination and degradation of B23 (31).
Because melanoma-associated ARF mutants have altered nucleolar localization (11) and ARF can mobilize out of the nucleolus in response to stress signals (14,33), we sought to refine the functional role of p14 ARF . We now show that endogenous p14 ARF and B23 fractionate with the nucleolar 60 S preribosomal particle. Moreover, p14 ARF and B23 interact within the 60 S complex in a rRNA-independent manner. p14 ARF does not significantly alter the amount of nuclear preribosomal particles but inhibits polysome formation and retards protein translation. We propose that the p14 ARF tumor suppressor functions to integrate cell growth and cell cycle progression with nucleolar ARF regulating ribosome function, whereas nucleoplasmic ARF responds to cellular stress by activating p53dependent growth arrest.
The U2OS_ARF, WMM1175_ARF, or WMM1175_p16 INK4a cell clones carrying the stably integrated p14 ARF gene or p16 INK4a gene under IPTG-inducible expression control have been described previously (35,36). The NARF2-E6 cells were initially provided by Dr. Gordon Peters (Cancer Research UK London Research) and have been described elsewhere (28). Briefly, these cells express an IPTG-inducible form of human p14 ARF and constitutively accumulate human papillomavirus E6 protein, which promotes p53 degradation. Stable cell clones were seeded 24 h prior to induction in the absence of antibiotics and were induced with 1-5 mM IPTG.
Indirect Immunofluorescence-Cultured cells (1 ϫ 10 5 ) seeded on coverslips in six-well plates were washed in phosphate-buffered saline and fixed in 3.7% formaldehyde. The cells were immunostained for 50 min with primary antibodies followed by a 50-min exposure to Texas Red-or FITC-conjugated secondary IgG (Sigma).
Cell Cycle Analysis-The cells were fixed in 70% ethanol at 4°C for at least 1 h, washed in phosphate-buffered saline, and stained with propidium iodide (50 ng/l) containing RNase A (50 ng/l). DNA content from at least 2000 cells was analyzed using ModFIT software.
Sucrose Density Gradient Fractionation-Cytoplasmic ribosomes were isolated as previously described (38), except that the lysis buffer contained only 0.2% Nonidet P-40. The lysates were centrifuged at 10,000 rpm for 10 min, and postmitochondrial supernatant was usually layered on 10 -45% or 10 -25% (w/w) sucrose density gradients in 10 mM Tris-HCl, pH 7.2, 60 mM KCl, 10 mM MgCl 2 , 1 mM dithiothreitol, 0.1 mg of heparin/ ml. To disrupt the 80 S cytoplasmic ribosome, MgCl 2 was omitted during lysis, and the lysates were separated in 5-25% (w/w) sucrose gradients in 10 mM Tris-HCl, pH 7.2, 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 0.1 mg of heparin/ml. The gradients were centrifuged at 36,000 rpm for 2 h 40 min at 5°C in a Beckman SW41Ti rotor and fractionated through a Bio-Rad EM-1 UV monitor for continuous measurement of the absorbance at 254 nm.
Isolation of nuclear extracts was performed as described (38). Nuclear proteins (1-1.5 mg) were overlaid on 15-30% (w/w) sucrose gradients in 25 mM Tris, pH 7.5, 100 mM KCl, 1 mM dithiothreitol, 2 mM EDTA and centrifuged at 38,000 rpm for 3 h 20 min at 5°C in a Beckman SW41Ti rotor. Where indicated, the nuclear extract was treated with RNase A (100 g/ml) for 10 min at room temperature. The gradients were analyzed as above. The proteins were precipitated from the fractions with trichloroacetic acid prior to Western blotting, and RNA was extracted from fractions using TRI-reagent LS (Sigma).
Immunoprecipitations-For immunoprecipitation analysis, U20S_ARF cells were left untreated or exposed to 1 mM IPTG for up to 96 h. Nuclear protein extracts (1.5 mg) were fractionated on sucrose gradients as detailed previously. Immunoprecipitations were performed overnight using 1 g of antibody chemically adsorbed to M450-tosylactivated DYNALbeads (DYNAL) as described by the manufacturer. Immunoprecipitates were washed four times with NET2 (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Nonidet P-40 containing protease inhibitors (Roche Applied Science)) lysis buffer, resolved using SDS-PAGE, and detected by immunoblot analysis. Extraction of RNA post-immunoprecipitation was performed using TRIreagent LS (Sigma).
Biochemical Fractionation-Cultured cells were permeabilized in buffer A (10 mM HEPES, pH 7.9, 15 mM KCl, 2 mM MgCl 2 , 0.1 mM EDTA, and protease inhibitors (Roche Applied Science)) for 10 min and then supplemented with 0.2% Nonidet P-40. The cytosol fraction was collected, and the cells were washed in buffer A prior to solubilization using equal volumes of buffer B (50 mM HEPES, pH 7.9, 50 mM KCl, 0.1 mM EDTA, and protease inhibitors) and buffer C (50 mM HEPES, pH 7.9, 0.8 M KCl, 0.1 mM EDTA, and protease inhibitors). The soluble nuclear fractions were collected after centrifugation. Soluble nuclear and cytosolic fractions derived from equal cell numbers were analyzed by immunoblotting.
Protein Synthesis Measurements-The cells (1 ϫ 10 4 ) were plated in methionine-free, Dulbecco's modified Eagle's medium in 96-well FLASH plates (PerkinElmer Life Sciences). Approximately 24 h post-seeding, the medium was replaced p14 ARF and Ribosomes DECEMBER

RESULTS
Impact of Induced p14 ARF Expression-To evaluate the influence of p14 ARF accumulation on cell proliferation, the U20S osteosarcoma cell line, which is wild type for both p53 and pRb (39), was engineered to express wild type p14 ARF . In this U20S_ARF cell line, p14 ARF expression was induced with 1 mM IPTG. Three days post-induction these cells behaved as expected, with strong nucleolar expression of p14 ARF (Fig. 1, A and B), increased p53 levels, increased expression of the p53transcriptional target p21 Waf1 , decreased levels of the transcription factor E2F-1 (40) (Fig. 1B), and potent G 1 phase cell cycle arrest (Fig. 1C). In addition, although p14 ARF has been reported to induce the degradation of B23 (31) and inhibit rRNA processing (29) (which would limit the production of mature rRNAs), we found that the presence of p14 ARF had no detectable effect on the accumulation of B23 (Fig. 1B), on the steady-state levels or processing of mature 28 and 18 S rRNA species (Fig. 1, D and E), on nucleolar transcription activity when RNA polymerase I transcription was visualized by pulse labeling with bromouridine and immunostaining with anti-bromodeoxyuridine antibody (Fig. 1F), or on cell size (Fig. 1G).
p14 ARF Associates with the 60 S Preribosomal Particles-The association of tandem affinity purification-tagged murine p19 ARF with ribosomal proteins and the fractionation of p19 ARF in large nuclear multiprotein complexes (2-5 MDa) (30) suggested that the ARF protein associates with preribosomal particles within the nucleolus. To examine this possibility, nuclear extracts derived from Saos 2 osteosarcoma cells, which express endogenous p14 ARF , were subjected to sucrose gradient centrifugation to separate preribosomal particles. Under these conditions three detectable peaks were resolved, and assignment of 40 and 60 S preribosomal subunits was confirmed by resolving the extracted RNA species and Western immunoblotting for ribosomal proteins, including antibodies of the small ribosomal subunit, S7 and the large ribosomal subunit, L28, and L37. Western blot detection of endogenous p14 ARF in Saos 2 cells confirmed that the majority of p14 ARF cosedimented with the 60 S preribosome particles ( Fig. 2A). Likewise, sucrose density gradient ultracentrifugation of IPTG-treated U20S_ARF nuclear cell extracts followed by Western immunodetection confirmed that a large proportion of the induced p14 ARF cosedimented with the 60 S preribosome particles (Fig. 2, B and C).
To verify that p14 ARF is a component of preribosomes, nuclear Saos 2 extracts were either left untreated or treated with RNase A before being subjected to fractionation on a sucrose gradient. As expected, treatment with RNase A disrupted ribosomal particles and resulted in the loss of preribosome peaks, and ribosomal protein S7, B23, and p14 ARF remained at the top of the sucrose gradient (Fig. 2D).
p14 ARF Interacts with B23 in the 60 S Preribosome Particles-To investigate whether p14 ARF is bound to known binding partners within preribosomes, we examined the sedimentation behavior of hdm2, p53, and B23. As shown in Fig. 2, a significant pool of B23 cosedimented with p14 ARF within the 60 S preribo-FIGURE 1. Accumulation of p14 ARF in U20S_ARF cells promotes p53 accumulation and cell cycle arrest but no detectable changes in rRNA synthesis or cell size. U20S_ARF cells were treated with 1 mM IPTG (ϩ) or left untreated (Ϫ) for 72 h, unless otherwise indicated. A, the subcellular distribution of p14 ARF and p53 was analyzed by immunofluorescence. LM, light microscopy. B, total proteins extracted from IPTG-treated or untreated cells were separated using SDS-PAGE and immunoblotted for the indicated proteins. C, the cell cycle distribution of induced and uninduced U20S_ARF was determined using propidium iodide staining. D, total RNA extracted from equal numbers of U20S_ARF cells was separated on a formaldehyde-containing agarose gel and visualized by ethidium bromide staining. E, following IPTG treatment (24 h), RNA was pulse labeled with [methyl-3 H]methionine for 30 min and chased in nonradioactive medium for the times indicated. Eight thousand 3 H cpm were resolved on agarose gel, transferred to a membrane, and visualized by fluorography. Positions of rRNAs and precursors are indicated on the left. F, permeabilized U20S_ARF cells, induced with 1 mM IPTG and showing a p14 ARF positive and negative cell, were pulse-labeled with bromouridine for 30 min before fixation to determine location of p14 ARF and ongoing RNA transcription. G, flow cytometric forward scatter profiles of uninduced and induced U20S_ARF cells.

p14 ARF and Ribosomes
some in both Saos 2 and induced U20S_ARF cells. It is worth noting that B23 also cosedimented with the 60 S preribosomal particles in the absence of p14 ARF (Fig. 2B). In contrast, hdm2 and its ubiquitination target, p53 were restricted to smaller molecular weight complexes near the top of the sucrose gradients, and this sedimentation behavior was not altered by p14 ARF (Fig. 2, B and C).
Considering that B23 and p14 ARF cosediment with the pre-60 S particle, we investigated whether these proteins are complexed within the preribosome. Nuclear extracts derived from IPTG-induced and uninduced U20S_ARF cells were separated using sucrose gradient fractionation. The fractions corresponding to the top of the column (containing small molecular weight complexes), the 40 S preribosome, and the 60 S preribosome were collected. The sucrose fractions (top, 40 and 60 S) were precipitated with antibodies to p14 ARF , and recovered immune complexes were denatured, separated on polyacrylamide gels and blotted with antibodies to B23, hdm2, and p53. As expected, the preribosomal fraction of p14 ARF was found complexed to B23 (Fig.  3A, lower panel) and not bound to hdm2 or p53 (data not shown).
We next examined whether the ARF-B23 interaction within the preribosome particles occurs independently of rRNA, because both proteins bind the 5.8 S rRNA found in the 60 S ribosome (29,30). Expression of p14 ARF was induced in U20S_ARF cells with 1 mM IPTG for 96 h, and nuclear extracts were prepared and separated using sucrose density centrifugation. The fractions corresponding to the 60 S preribosome were collected and treated with RNase A. Extraction of the 60 S fraction post-immunoprecipitation with the RNA isolation reagent TRI-Reagent LS confirmed that no detectable RNA was present in the RNase A-treated 60 S preribosomal fractions (Fig. 3B, upper  panel). In the absence of RNA, endogenous B23 was still found in association with p14 ARF in the 60 S preribosomal fraction (Fig. 3B,  lower panel).
p14 ARF Expression Does Not Influence Preribosome Production-To examine the influence of p14 ARF accumulation on nucleolar ribosome assembly, U20S_ARF cells were grown in the absence or presence of 1 mM IPTG for up to 96 h. Equal amounts of nuclear proteins were separated on 15-30% sucrose density gradients. Analysis of 40 and 60 S preribosomal particles revealed no significant changes in the ratio of 60 to 40 S preribosomes subunits at 48, 72, and 96 h post-induction (Fig. 2, A and B). In particular, the peak height ratio of 60 to 40 S subunits in untreated and treated U20S_ARF cells, 96 h post-induction, was estimated to be 1.6 Ϯ 0.1 and 1.6 Ϯ 0.1, respectively (the means of three independent experiments).  (bottom panel). B, the inducible U20S_ARF cell line was grown in the absence (Ϫ) of IPTG for 48 h. Preribosomes were fractionated and analyzed as detailed above. C, the inducible U20S_ARF cell line was grown in the presence (ϩ) of IPTG for 48 h. Preribosomes were fractionated and analyzed as detailed above. D, nuclear Saos 2 extracts were either treated with RNase A or left untreated and analyzed as detailed above. DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 p14 ARF Expression Inhibits Translation Initiation-The association of p14 ARF with preribosomal particles suggested that this tumor suppressor protein might influence cytoplasmic ribosome assembly and/or function. When equal numbers of U20S_ARF cells were extracted and the cytoplasmic lysates separated on 10 -25% (to focus on monosome particles) or 10 -45% (allows polysome separation) sucrose density gradients, a shift in the distribution of ribosomes to smaller polysomes and larger pools of 80 S monosomes was consistently observed (Fig. 4, A and B). To determine whether the ARF-induced increase in monosome 80S particles was due to reduced polysome levels, we separated all 80 S subunits, including polysomes, into their 40 and 60 S components by preparing FIGURE 3. p14 ARF and B23 interact and associate with 60 S preribosomes. A, U20S_ARF cells were grown in the presence or absence of 1 mM IPTG for 48 h, and preribosomes were fractionated on 10 -30% sucrose density gradients (upper panel, induced U20S_ARF fractionation shown). Fractions corresponding to small molecular weight complexes (top of gradient), 40 S preribosomes and 60 S preribosomes were immunoprecipitated with a monoclonal ARF or isotype-matched antibody (IgG). The immunoprecipitates were analyzed for the presence of B23 by Western blotting (lower panel). B, sucrose gradient fractions corresponding to 60 S preribosomes derived from IPTG-induced and uninduced U20S_ARF cells were treated with RNase A for 30 min at room temperature. The treated fractions were immunoprecipitated with a monoclonal ARF or isotype-matched antibody (IgG). To confirm that rRNA was effectively degraded by RNase treatment, the post-immunoprecipitation lysates were extracted using TRI-reagent LS. Extracted RNA and total cell RNA were resolved in a formaldehyde-containing agarose gel and stained with ethidium bromide (upper panel). Immunoprecipitates (IP) were analyzed for the presence of B23 by Western blotting (lower panel).

FIGURE 4. Expression of p14 ARF increases formation of inactive 80 S monosomes and reduces polysome formation.
U20S_ARF cells were grown in the presence (ϩ) or absence (Ϫ) of 1 mM IPTG for 72 h. A, cytoplasmic ribosomes were fractionated on 10 -25% sucrose density gradients, to separate monosomes and analyzed with continuous monitoring of absorbance at 254 nm. B, cytoplasmic ribosomes were fractionated on 10 -45% sucrose density gradients to allow for polysome separation and analyzed with continuous monitoring of absorbance at 254 nm. C, cytoplasmic lysates derived from induced and uninduced U20S_ARF cells were prepared in the absence of MgCl 2 to separate 80 S ribosomes into their 40 and 60 S subunits. The ribosomes were fractionated on 5-25% sucrose density gradients with continuous monitoring of absorbance at 254 nm. D, cytoplasmic lysates derived from induced and uninduced U20S_ARF cells were prepared in the presence of 0.8 M NaCl to dissociate 80 S couples, which consist of 40 and 60 S subunits joined in the absence of mRNA. The ribosomes were fractionated on 10 -20% sucrose density gradients with continuous monitoring of absorbance at 254 nm.

p14 ARF and Ribosomes
cytoplasmic lysates in the absence of MgCl 2 . The disruption of 80 S ribosomes resulted in similar levels of 40 and 60 S subunits in the induced and uninduced U20S_ARF cells (Fig. 4C). Thus, p14 ARF did not alter the accumulation of 40 and 60 S ribosome subunits within the cytoplasm but altered the complexes that they produced. In particular, p14 ARF expression caused a significant increase in the proportion of monomeric 80 S subunits (Fig. 4A). This was accompanied by a significant reduction in protein translation; U20S_ARF cells induced to express p14 ARF for ϳ72 h incorporated 26 Ϯ 2% less [ 3 H]methionine than the uninduced control cells (Fig. 5A). Further, these ARF-expressing cells accumulated less of the translationally regulated c-Myc and cyclin D1 proteins (41) (Fig. 5B).
The rate of synthesis of any given protein is determined primarily by the level of translation initiation, and the increase in monosome 80 S subunits in p14 ARF -expressing cells was indicative of a block in the initiation phase of translation. The appearance of a large population of nontranslating 80 S couples, consisting of 40 and 60 S subunits joined in the absence of mRNA, is typical of an initiation defect. High levels of salt can readily dissociate these inactive 80 S particles (42), and we examined the effect of including 0.8 M NaCl in the sucrose gradient. As expected, dissociation of inactive couples led to an increase in the levels of cytoplasmic 40 and 60 S ribosome subunits in the p14 ARF -expressing U20S_ARF cells. Thus, inactive 80S monosomes are dramatically increased in the presence of p14 ARF (Fig. 4A) and constitute a major proportion of the ribosome population in the ARF-expressing cells (Fig. 4D).
To gain some insight into the mechanism of this translation defect, the accumulation of ␣-subunit of the eukaryotic initiation factor 2 (eIF2) was examined. eIF2 facilitates binding of the initiator tRNA to the 40 S subunit during translation initiation. Translation can be inhibited by the rapid phosphorylation of the ␣-subunit (reviewed in Ref. 43), by limiting its expression and accumulation (44), or by specific caspase-mediated eIF2␣ cleavage during apoptosis (45). Further, phosphorylation of eIF␣ has been associated with an increase in the formation of monosome 80 S subunits (46). In this study we found that ARFexpressing U20S cells showed a marked reduction in the accumulation of the eIF2␣ subunit (Fig. 5B).
The observed changes in cytoplasmic ribosomes, in particular the increase in the monomers, was not an artifact of IPTG exposure, because addition of IPTG to the parental U20S_lac17 cell line, which expresses the lac repressor, did not alter the ribosomal subunit profile (data not shown). To investigate whether p14 ARF -induced effects on polysome production were a result of cell cycle arrest and p53-dependent, we utilized the WMM1175_ARF melanoma and NARF2-E6 osteosarcoma cell lines. Both cell lines express an IPTG-inducible form of p14 ARF but are functionally null for p53 function (Fig. 6A). The induction of p14 ARF in the WMM1175_ARF and NARF2-E6 cells did not induce growth arrest (Fig. 6A) but did promote an increase in 80 S monomeric ribosomes and a reduction in polysomes (Fig. 6, B and C). This ARF-associated decrease in the polysome formation in the WMM1175_ARF cell line is substantial, considering that only 40% of the WMM1175_ARF cells express detectable levels of ARF upon IPTG induction.
To further examine the relationship between cell cycle arrest and inhibition of polysome production, we promoted G 1 cell cycle arrest in the WMM1175_p16 INK4a cell line by inducing the expression of the cyclin-dependent kinase inhibitor p16 INK4a with 5 mM IPTG for 72 h (36). Although, p16 INK4a expression induced potent G 1 cell cycle arrest (Fig. 7A), it led to an increase in cell size (Fig. 7B) that was accompanied by an increase in the level of cytoplasmic ribosomes and polysome production (Fig. 7C).
Considering that p14 ARF reduced cytoplasmic ribosome function, we also examined whether p14 ARF cosedimented with ribosomes within the nucleolus and cytoplasm. We found that the majority of p14 ARF in Saos 2 cells and induced U20S_ARF cells occurred in the nuclear fraction, with no detectable cytoplasmic p14 ARF component (Fig. 8).

DISCUSSION
The ARF tumor suppressor plays a central role in limiting cell cycle progression in response to hyperproliferative signals. Cells exposed to oncogenes, such as E2F-1, Ras, or Myc (47)(48)(49), accumulate nucleolar ARF, and yet its core function of stabilizing p53 requires its association with hdm2 in the nucleoplasm (21). In this study we sought to explore the interaction of ARF with the nucleolus and have shown that p14 ARF is a component of the 60 S preribosome, where it interacts with B23 independently of rRNA. The association of B23 with the 60 S particle is not unexpected, because this protein binds the 28 S rRNA, a major component of the pre-60 S ribosome (50). We also found that ARF did not influence the production of the preribosomes or the steady-state levels of cytoplasmic ribosome subunits but inhibited ribosome function. The impact of p14 ARF on ribosome activity may involve the translation initiation factor eIF2␣; ARF-expressing U20S cells accumulated low levels of this subunit. Reduced expression of eIF2␣ in quiescent cells correlates with low levels of protein synthesis (44), whereas induction of eIF2␣ expression via oncogenes such as c-Myc contributes to the pronounced stimulation of protein production (51). eIF2␣ is also involved in neoplastic transformation when its function is up-regulated (reviewed in Ref. 52). We are currently investigating whether repression of eIF2␣ in response to p14 ARF expression involves the c-Myc transcrip-  DECEMBER 8, 2006 • VOLUME 281 • NUMBER 49 tion factor; p14 ARF has been shown to interact with and inhibit the transcriptional activity of the c-Myc oncogene (53,54).

p14 ARF and Ribosomes
Our findings extend and help clarify recent reports that describe the ARF-B23 association and thus implicate ARF in the regulation of protein synthesis. Data demonstrating that the murine homologue of p14 ARF , p19 ARF , promoted the degradation of B23 (31) and inhibited rRNA processing is not supported by our findings with p14 ARF or indeed with other studies (21,55). We found that B23 and ARF associate within the nucleolar pre-60 S ribosome, and it has been reported that nucleolar B23 is resistant to ARF-induced degradation (31). It has also been suggested that the interaction of p19 ARF and B23 retains both proteins in the nucleolus, effectively impeding the nucleocytoplasmic shuttling of B23 to induce growth arrest (55) but also blocking p19 ARF -mediated p53 activation (21). Importantly, hdm2 and B23 compete for ARF binding, and thus silencing of B23 expression enhanced the ARF-mdm2 association within the nucleoplasm and induced p53-activated growth arrest (19). The implications of these findings are that p19 ARF directly accesses ribosome function through its nucleolar association with B23 and that it regulates the p53 cell cycle pathway via its nucleoplasmic interaction with hdm2. This model is supported by our studies with human p14 ARF . We have shown that the predominantly nucleolar p14 ARF was associated with B23 within the 60 S preribosome subunit, and this interaction did not involve hdm2 or p53, both of which were excluded from nucleoli and the preribosomes. Further, the ARF-mediated inhibition of protein translation in vivo supports the notion that ARF modulates ribosome function. We propose that ARF is a highly mobile protein, and as it accumulates in response to oncogenic activity, the increased nucleoplasmic fraction complexes hdm2, in preference to B23 (55), and activates p53-mediated growth arrest. Likewise, a fraction of nucleolar ARF, possibly a fraction tethered to B23, may impact ribosomes within the nucleolus. This work extends the already diverse functional repertoire of ARF to that of effects on protein translation; thus, ARF may serve to coordinate the regulation of cellular growth with that of proliferation. When these programs become uncoupled, as reported in T-ALL cells reconstituted to express p16 INK4a (56), cells undergo proliferative arrest but continue to grow and differentiate.
There is strong precedent for the involvement of nucleolar proteins in the regulation of cell proliferation and cell growth. Nucleolar release of the ribosomal proteins L23, L11, and L5 promotes their interaction with hdm2, resulting in hdm2 inactivation and stabilization of p53. Similarly, the nucleolar protein Bop1 inhibits rRNA formation and also activates p53-dependent cell cycle arrest (38,57). Analogous to ARF, nucleolar B23 contributes to multiple steps in ribosome biogenesis (58), and its nucleoplasmic redistribution activates p53 by inhibiting hdm2 (27). In all cases, these nucleolar regulators can rapidly halt cell cycle progression by leaching out of the nucleolus and signaling through the p53 pathway and also regulate the activity of ribosomes. We found that p14 ARF promoted p53 activity and cell cycle arrest within 16 h of induction (the percentage of S phase inhibition in U20S_ARF cells induced for 16 h was 47%), whereas the ability of ARF to limit translation was not observed until 3 days after it was induced. This is consistent with the slower growth inhibitory impact of p14 ARF expression in p53-null cells (59). Although p14 ARF can alter ribosome function in the absence of p53, it is likely that inhibition of translation is enhanced in the presence of both p14 ARF and p53, because both tumor suppressors independently inhibit ribosome function (60). The impact of p53 on translation is minor, however, and requires the inter-FIGURE 6. p14 ARF reduces polysome formation in a p53-indpendent manner. The inducible p53-null WMM1175_ARF and NARF2-E6 cell lines were grown in the presence (ϩ) or absence (Ϫ) of IPTG for 72 h. A, cell lysates were separated on 12% SDS-polyacrylamide gels and immunoblotted for the indicated proteins. Cell cycle distribution of cells either untreated or treated for 72 h with IPTG using propidium iodide staining. B, cytoplasmic ribosomes were fractionated on 10 -45% sucrose density gradients for the WMM1175_ARF cells with continuous monitoring of absorbance at 254 nm. C, cytoplasmic ribosomes were fractionated on 15-45% sucrose density gradients for the NARF2-E6 cells with continuous monitoring of absorbance at 254 nm. The vertical arrows indicate the 80 S monomer peaks.

p14 ARF and Ribosomes
action of p53 with polyribosomes (61), an association that we did not observe in ARF-expressing U20S cells (data not shown).
The increasing number of nucleolar proteins involved in regulating both ribosome biogenesis and p53 function emphasizes the pivotal role of the nucleolus in synchronizing cell cycle progression and cell growth. The majority of stress signals disrupt the nucleolus, causing the release of many proteins, including p14 ARF , B23, L11, L23, and L5, which interact with and inhibit hdm2 and thus activate p53. Despite the many functional similarities within this group of proteins, there is strong evidence of functional diversity. For instance, B23 appears necessary for UV-induced activation of p53 (27), L23 is important for actinomycin D-caused p53 induction (25), and p14 ARF , although not required for UV-or actinomycin-induced p53, is essential for the activation of p53 in response to oncogenes. In fact, the role of p14 ARF in response to hyperproliferative oncogenic signals may account for its ribosome regulatory function. Expression of p14 ARF is induced by the same oncogenes (Myc and Ras) that rapidly increase rRNA transcription rates (62,63). Thus, in cells expressing activated oncogenes, the increased expression of p14 ARF would dampen the aberrant up-regulation of ribosome subunits. In contrast, in the absence of p14 ARF , oncogenic stimulation would promote increased ribosome biogenesis, enhanced translation rates, and ultimately transformation.  . p14 ARF protein accumulates in the nuclear fraction. Saos 2 cells, uninduced U20S_ARF and IPTG-induced U20S_ARF cells were partitioned into Nonidet P-40 soluble cytosol (C lanes) and soluble nuclear (N lanes) fractions. Proteins corresponding to equal cell numbers were fractionated using SDS-PAGE and immunoblotted for the indicated proteins. The effective fractionation of cells was verified using antibodies against nuclear topoisomerase II and cytoplasmic tubulin.