A Major Functional Difference between the Mouse and Human ARF Tumor Suppressor Proteins*

Suppression of tumorigenesis is considerably more stringent in the human than in the much shorter lived mouse species, and the reasons for this difference are poorly understood. We investigated functional differences in the control of the ARF (alternative reading frame) protein that acts upstream of p53 and is encoded along with p16 INK4a at a major tumor suppressor locus in both the human and mouse genomes. The mouse and human ARF proteins are substantially divergent at their carboxyl termini. We have shown that the mouse ARF protein (p19ARF) interacts with Pex19p in the cell cytoplasm leading to its nuclear exclusion and repression of its p53 activation function. The human ARF protein (p14ARF) is substantially smaller than its mouse coun-terpart and is not subject to this functional inactivation by Pex19p. In an identical cellular background, ribozymes directed against Pex19p enhanced p19ARF- ) and antisense (5 (cid:1) -gtc gac gca cct aga gag agg-3 (cid:1) ) Pex19p-specific primers with EcoR I and Sal I sites, respectively. The PCR amplification (94 °C for 30 s, 55 °C for 30 s, and 72 °C for 3 min) product was purified and sequentially ligated to pGEM-T easy (Pro-mega), pODB8 and pACT2 (yeast two-hybrid vectors (48)), pVP-16 (mammalian two-hybrid vector, CLONTECH), pEGFPC1 (mammalian

Pex19p is a farnesylated cytosolic protein that acts as a soluble receptor or chaperone for targeting of peroxisomal membrane proteins (37). It plays an important role in peroxisomal biogenesis, membrane assembly, and stabilization (38 -40). We previously showed that p19ARF-Pex19p interactions in the cytoplasm result in cytoplasmic retention of p19ARF and functional dampening of its p53 activation function (36). This effect on p53 is in accord with the proposed involvement of nuclear p19ARF-MDM2 interactions in restraining MDM2-mediated degradation of p53.
Recent studies have shown that ARF function involves complex feedback mechanisms (41,42). Its expression is regulated by E2F, and thus binding of hypophosphorylated pRB to E2F inhibits ARF expression and its downstream p53 activation. This pathway thus provides a regulatory link between pRb and p53 pathways; inactivation of pRb by phosphorylation and release of E2F from pRb-E2F complexes leads to an activation of ARF expression resulting in the stabilization and functional activation of p53 (5,43,44). Activation of the ARF-p53-p21 WAF1 pathway can in turn restrict phosphorylation-mediated inhibition of pRB function by inhibiting cyclin-cyclin D-dependent kinase activity.
Functional regulation of ARF is critical for cell cycle control in response to a variety of cellular and environmental signals. The mouse and human ARF proteins share only a limited homology at the cDNA and protein levels (1,(45)(46)(47), and the functional relevance of this genetic divergence is unknown. We report here that a result of this difference is that the human ARF protein is not inactivated by Pex19p. This may contribute to the more stringent control of cellular senescence and tumor suppression in human cells.

MATERIALS AND METHODS
Plasmid Construction-Full-length mouse and human Pex19p cDNAs were cloned from mouse and human testis by RT-PCR 1 using mouse sense (5Ј-gaa ttc atg gcg gct gct gag gaa ggt-3Ј) and antisense (5Ј-gtc gac tca cat gat cag aca ctg ttc-3Ј) and human sense (5Ј-gaa ttc atg gcc gcc gct gag-3Ј) and antisense (5Ј-gtc gac gca cct aga gag agg-3Ј) Pex19p-specific primers with EcoRI and SalI sites, respectively. The PCR amplification (94°C for 30 s, 55°C for 30 s, and 72°C for 3 min) product was purified and sequentially ligated to pGEM-T easy (Promega), pODB8 and pACT2 (yeast two-hybrid vectors (48)), pVP-16 (mammalian two-hybrid vector, CLONTECH), pEGFPC1 (mammalian * This work was supported in part by a Hougateki research grant from the National Institute of Advanced Industrial Science and Technology (to S. C. K.) and by a carcinogenesis fellowship from the Cancer Council of New South Wales (to R. R. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Yeast Two-hybrid Interactions-Yeast reporter strains PJ69/2A and Y187 (Trp Ϫ /Leu Ϫ /His Ϫ Ade Ϫ ) (48) were transformed with pODB8 plasmid constructs encoding full-length p19ARF and its various deletion mutants or p14ARF. The selected cells were secondarily transformed with the pACT-2/mPex19p or hPex19p constructs. Double transformants that grew on Trp Ϫ /Leu Ϫ /His Ϫ /Ade Ϫ selection medium were analyzed for the presence of ARF and Pex19p sequences by PCR and were assayed for ␤-galactosidase reporter activity. Cell extracts were prepared using standard conditions, and enzyme activity was determined using the GAL-Tropix kit according to the manufacturer's protocol (Tropix Inc.).
Cell Culture and Transfections-Mouse embryonic fibroblasts, NIH 3T3, and monkey kidney cells (COS 7) were cultured in Dulbecco's modified Eagle's minimal essential medium supplemented with 10% fetal bovine serum. Transfections were performed using Lipo-fectAMINE (Invitrogen). Typically, 1 g of plasmid DNA was used per well in a 24-well dish, and 3 g was used per 6-cm dish. For immunostaining, cells were plated on glass coverslips and transfected at 60% confluency. Cells were fixed with methanol:acetone (1:1) at the indicated time intervals following transfections and immunostained as described below. NIH 3T3 cells transfected stably with a construct containing p19ARF cDNA driven by the metallothionein promoter were a kind gift from J. Kato (52). Expression of p19ARF was induced by supplementation of growth medium with 100 M ZnSO 4 . Cells transfected with ribozymes were selected in puromycin-supplemented medium (5 g/ml for 2 days followed by 0.5 g/ml for the next 2 days). Expression of ribozymes was detected by RT-PCR, and the effect on Pex19p expression was analyzed by Western blotting.
Mammalian Two-hybrid Analysis-COS 7 cells were seeded at 50 -60% confluence in 24-well plates and transfected with 1 g of DNA containing pG5-reporter plasmid, pM/ARF, pVP16/Pex19p, and pM or pVP16 control vectors as indicated in the relevant figure legends. 5-7 h after transfection, cells were refed with fresh medium and were lysed in universal lysis buffer (Promega) after 48 h. Luciferase activity was measured by using the dual luciferase reporter assay system (Promega). Results presented are the means of at least three transfections.
In Vivo Coimmunoprecipitation-Cells transfected with plasmids encoding Myc-tagged ARF and GFP-Pex19p fusion proteins were lysed in Nonidet P-40 lysis buffer. For immunoprecipitation of Myc-tagged ARF proteins, lysate containing 400 g of protein was incubated with a polyclonal anti-Myc antibody (Santa Cruz SC-789) at 4°C for 1-2 h. Immunocomplexes were separated by incubation with Protein A/G-Sepharose, and Western blotting was performed with a monoclonal anti-GFP antibody (CLONTECH 8362-1) or an anti-Xpress antibody FIG. 1. p19ARF, but not p14ARF, interacts with Pex19p in twohybrid assays. A, yeast reporter strain Y187 was transformed with expression plasmid constructs encoding Gal4 DNA binding domain-ARF (as indicated) fusion proteins and Gal4 DNA activation domain-Pex19p fusion protein. Double transformants that grew on selection plates were subjected to two-hybrid ␤-galactosidase reporter activity filter assay. Yeast colonies grown on filter paper (Whatman no. 3) were lysed by dipping the filter paper in liquid nitrogen. ␤-Galactosidase activity (development of a blue color) was monitored by incubating the filter in X-gal (5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside) for 3-5 h at 37°C. Interactions between the two proteins were observed as the blue color of the colonies, and lack of interactions was marked by white colonies. The deletion constructs of ARF that lacked the carboxyl terminus region did not interact with Pex19p. p14ARF was also negative. a.a., amino acids. B, COS 7 cells were transfected with mammalian two-hybrid plasmids (pG5-reporter plasmid, pM-ARF, pVP16-Pex19p, pM, and pVP16 as indicated). Following 48 h of transfections, cells were assayed for luciferase (Luc) activity (dual luciferase reporter assay system, Promega). pCMV-thymidine kinase (TK) renilla luciferase reporter was used as an internal control for transfection efficiency. Results presented are the mean of three transfections. Interactions were positive for p19ARF but negative for p14ARF.  (panel b, lane 3). A faint band (close to the size of Pex19p) cross-reacting to GFP antibody was detected in lanes 1-3. GFP by itself showed no coimmunoprecipitation with p19ARF- Myc (panel b, lanes 4 -6). Input signals (from 10% of the lysates) for GFP- Pex19p (panel a, lanes 1-3) and GFP (panel a, lanes 4 -6) are shown.
(Invitrogen R910-25) as indicated and a horseradish peroxidase-conjugated secondary antibody (ECL kit, Amersham Biosciences) using standard procedures and detection by ECL chemiluminescence.
Immunostaining-Cells grown on glass coverslips were fixed by incubation with prechilled methanol/acetone (1:1) for 5 min on ice. These were washed with PBS and blocked with 0.2% bovine serum albumin in PBS for 20 min. Cells were then incubated in primary antibody (anti-Myc, in blocking buffer) for 1-2 h. Stained cells were visualized by secondary staining with Alexa Fluor™ 488 goat anti-rabbit IgG conjugate (Molecular Probes). After six washes in PBS with 0.1% Triton X-100, cells were overlaid with a coverslip with Fluoromount (Difco). The cells were examined on a Zeiss microscope with epifluorescence optics or a Fluoview confocal laser-scanning microscope (Olympus, Tokyo, Japan).
Colony-forming Assays-NIH 3T3 cells were stably transfected with expression plasmids encoding p19ARF, p14ARF, or p19ARF(d-C41) (a deletion mutant of p19ARF lacking the carboxyl-terminal 41 amino acids) driven by the metallothionein promoter. Transfected cells were selected in G418-supplemented medium, plated in 10-cm dishes (500 cells/dish), induced for ARF proteins by the addition of 100 M ZnSO 4 , and examined for colony formation for 2 weeks.
Reporter Assays-NIH 3T3 cells stably transfected with a p53-responsive luciferase reporter plasmid, PG-13luc (kindly provided by Dr. Bert Vogelstein) (53) and expression plasmids encoding metal-inducible p19ARF or p14ARF proteins were transfected with Pex19p ribozymes. Transfected cells were selected by puromycin and were then induced for ARF expression for 24 -48 h. As a control, pRL-TK vector (Promega) was co-transfected in each assay to correct for variations in transfection efficiency. Cells were lysed and luciferase activity was measured by using the dual luciferase reporter assay system (Promega).

RESULTS AND DISCUSSION
To elucidate the Pex19p binding domain of p19ARF, ␤-galactosidase reporter assay (dependent on the interactions of two proteins) was performed on yeast cells transformed with Pex19p and full-length p19ARF and its various deletion mutants or p14ARF. The full p19ARF protein and deletion mutants that retain the carboxyl-terminal 41 amino acid residues were positive for interactions with Pex19p. Based on these data, the Pex19p binding domain of p19ARF was assigned to its carboxyl-terminal 41 amino acid residues (Fig. 1A, 129 -169  a.a.). The human ARF protein p14ARF is shorter than the mouse ARF protein, p19ARF, by 40 amino acids (43). Notably, p14ARF-Pex19p interactions were negative in the yeast twohybrid system. The mouse and human ARF proteins were also tested for interaction with Pex19p by a mammalian two-hybrid reporter assay (Fig. 1B). The ARF cDNAs were cloned in-frame with the GAL4 DNA binding domain and were expressed in cells along with the DNA activation domain-Pex19p fusion protein. In this assay system, luciferase reporter activity is dependent on the interactions of the DNA binding and activation domains. It detected interaction of Pex19p with p19ARF but not with p14ARF (Fig. 1B). Thus, the yeast and mammalian two-hybrid assays both suggested that p14ARF does not interact with Pex19p. These findings were confirmed by in vivo co-immunoprecipitation assays (Fig. 2). Whereas immunoprecipitation of p19ARF pulled down Pex19p, an equivalent immunoprecipitation of p14ARF did not (Fig. 2). It was noted that (i) p14ARF runs very close to the dye front on 4 -20% SDS-PAGE and (ii) although the amount of immunoprecipitated p14ARF was greater than that of p19ARF from an equal quantity of lysate (Fig. 2, compare lanes 1 and 2), there was no coimmunoprecipitation of Pex19p with p14ARF. This result strongly supported the two-hybrid assays. We therefore concluded that in contrast to p19ARF, p14ARF does not interact with Pex19p.
Because Pex19p was shown to sequester p19ARF in the cytoplasm, we compared the subcellular localization of exogenous Myc-tagged mouse and human ARF proteins in an identical cellular background (HeLa cells). In time course experiments, p19ARF (as detected by staining with anti-Myc antibody) localized first in the cytoplasm (Fig. 3A, a) and subsequently moved to the nucleus and then to the nucleolus (Fig.  3A, b-d). In contrast, p14ARF was visible in the nucleus even at the earliest time point (6 h) (Fig. 3B, b), and as expected, p19ARF, but not p14ARF, colocalized with Pex19p in the cytoplasm (Ref. 36 and data not shown). The cellular background (HeLa cells), the exogenous promoter driving the ARF expression construct, and the level of expression of the two proteins as detected by Western blotting of the transfected cells with anti-Myc antibody (Fig. 3C) were identical in these experiments. Therefore the most likely reason for the different behavior of It was shown previously that p19ARF-Pex19p interactions result in the dampening of p19ARF function (36). We next compared the biological effects of p19ARF, p14ARF, and p19ARF(d-C41) (a deletion mutant lacking the carboxyl-terminal 41 amino acids) on colony-forming assays. NIH 3T3 cells (which lack endogenous ARF protein (54)) were stably transfected with metal-inducible expression plasmids encoding the above ARF proteins. The selected colonies were analyzed for protein expression by Western blotting, and it was found that there were equal levels of expression of the transfected proteins (similar to the data shown in Fig. 3C). The G418-selected cells were assayed for colony-forming efficiency with or without ARF expression (induced by the addition of 100 M ZnSO 4 into the medium). Expression of p14ARF resulted in 84 -87% reduction in the colony-forming efficiency. The expression of p19ARF caused 45-50% reduction (Fig. 4). These results showed that the growth suppressor activity of p14ARF is much stronger than p19ARF. Interestingly, p19ARF(d-C41) had a stronger effect than full-length p19ARF. These results, together with the finding that Pex19p binds to the carboxyl-terminal 40 amino acids of p19ARF and retains it, but not p14ARF, in the cytoplasm (Fig. 3), suggest that p14ARF and p19ARF(d-C41) translocate more rapidly into the nucleus (Fig. 3) resulting in stronger growth suppressor activity.
We next constructed hammerhead ribozymes to target Pex19p expression in NIH 3T3 cells. Target sites flanking the 10 putative ribozyme cleavage sites (GUC, GUA, CUC, and CUA) in the 5Ј terminus of Pex19p cDNA sequence were selected. Putative structures of each of the target sites along with the ribozyme and the tRNA sequence (155 nucleotides) were predicted using a RNA software (Mulfold2 and LoopViewer) as shown in Fig. 5A. Four target sites (with cleavage sites at nucleotides 40, 108, 122, or 168) that showed at least 60% open structure when embedded in ribozyme and tRNA sequences (Fig. 5A) were selected for construction in the pPUR-KE vector as described (49 -51). To select effective ribozymes, NIH 3T3 cells stably expressing His-Max-tagged mouse Pex19p were first made. These were transfected with Pex19p target ribozymes (Fig. 5A). Expression of ribozymes was analyzed by RT-PCR (data not shown), and their effectiveness against Pex19p was analyzed by Western blotting with anti-Xpress antibody (Fig. 5B). Two of the four ribozymes (Rz-40 and Rz-122) reduced Pex19p expression level to nearly one-tenth of the control cells (Fig. 5B). Inactive versions of these ribozymes (change of nucleotide G 5 to A 5 within the catalytic domain of the ribozyme (51)) were made to analyze their specificity; these did not affect Pex19p expression (Fig. 5B). To elucidate the effect of Pex19p targeting on p19ARF or p14ARF activity, we performed p53-dependent luciferase reporter assays. NIH 3T3 cells stably transfected with a p53-dependent luciferase reporter plasmid and metal-inducible expression of p19ARF or p14ARF were used. Cells transfected with ribozymes were selected with puromycin, induced for ARF expression for 48 h, and analyzed by luciferase assay. As expected, p19ARF or p14ARF resulted in up-regulation of p53-dependent reporter activity (Fig. 5C). Coexpression of Pex19p ribozymes (Rz-40 and Rz-122) resulted in further enhancement of p19ARF-induced p53 activity by 40%; the inactive versions of these ribozymes were neutral (Fig. 5C). These ribozymes had no effect in the absence of p19ARF. Most notably, these ribozymes did not affect p14ARF-dependent p53 transcriptional activation function (Fig. 5C). Taken together, the results showed that Pex19p interacts with mouse ARF protein and inactivates its function; human ARF by lacking a Pex19p binding region escapes from such inactivation.
ARF is a major tumor suppressor and is involved in cellular senescence (6,11,13,14). It has long been known that there are substantial differences in regulation of cellular senescence in mouse and human cells and that mouse cells become immortalized much more readily than human cells. The underlying reasons and mechanisms for these differences are not yet understood. We have shown that the ARF protein, a major player in cellular senescence and tumor suppression, is controlled differently in mouse and human cells. Its activity is decreased because of its interactions with Pex19p in the cytoplasm of mouse but not of human cells. This difference may provide a mechanistic explanation of the more stringent imposition of senescence in human cells and their resistance to immortalization.