Down-regulation of Myc as a potential target for growth arrest induced by human polynucleotide phosphorylase (hPNPaseold-35) in human melanoma cells.

Terminal differentiation and senescence share several common properties, including irreversible cessation of growth and changes in gene expression profiles. To identify molecules that converge in both processes, an overlapping pathway screening was employed that identified old-35, which is human polynucleotide phosphorylase (hPNPaseold-35), a 3',5'-exoribonuclease. We previously demonstrated that hPNPaseold-35 is a type I interferon-inducible gene that is also induced in senescent fibroblasts. In vitro RNA degradation assays confirmed its exoribonuclease properties, and overexpression of hPNPaseold-35 resulted in growth suppression in HO-1 human melanoma cells. The present study examined the molecular mechanism of the growth-arresting property of hPNPaseold-35. When overexpressed by means of a replication-incompetent adenoviral vector (Ad.hPNPaseold-35), hPNPaseold-35 inhibited cell growth in all cell lines tested. Analysis of cell cycle revealed that infection of HO-1 cells with Ad.hPNPaseold-35 resulted in arrest in the G1 phase and eventually apoptosis accompanied by marked reduction in the S phase. Infection with Ad.hPNPaseold-35 resulted in reduction in expression of the c-myc mRNA and Myc protein and modulated the expression of proteins regulating G1 checkpoint and apoptosis. In vitro mRNA degradation assays revealed that hPNPaseOLD-35 degraded c-myc mRNA. Overexpression of Myc partially but significantly protected HO-1 cells from Ad.hPNPaseold-35-induced growth arrest, indicating that Myc down-regulation might directly mediate the growth-inhibitory properties of Ad.hPNPaseold-35. Inhibition of hPNPaseold-35 by an antisense approach provided partial but significant protection against interferon-beta-mediated growth inhibition, thus demonstrating the biological significance of hPNPaseold-35 in interferon action.

There are two contrasting endpoints in the life of a replicating cell. One involves the normal physiological processes of differentiation or senescence. The other is the pathological process of neoplastic transformation characterized by uncontrolled proliferation and de-differentiation. Treatment of HO-1 metastatic human melanoma cells with fibroblast interferon (IFN-␤) 1 and the protein kinase C activator mezerein (MEZ) induces irreversible growth arrest and terminal differentiation characterized by changes in cell morphology, increase in melanin synthesis, modifications in gene expression, and alterations in surface antigen expression (1)(2)(3)(4)(5). Replicative or cellular senescence, a process leading to irreversible arrest of cell division, was first described in cultures of human fibroblasts that lost the ability to divide upon continuous subcultures (6). Replicative senescence can result from telomere shortening linked with a DNA end-replication problem, overexpression of certain oncogenes, or tumor suppressor genes, or it can be stress-induced premature senescence after exposure to a variety of oxidative stresses or DNA damaging agents (for a review, see Ref. 7).
Terminal differentiation and cellular senescence share several common traits including irreversible growth arrest and changes in gene expression profiles. To understand the molecular and biochemical basis of the complex physiological changes associated with these phenomena, an overlapping pathway screen was used to identify genes displaying coordinated expression as a consequence of both processes (8). A temporally spaced terminally differentiated human melanoma subtracted cDNA library was screened with cDNAs derived from senescent progeroid fibroblast cells. This led to the identification of old-35, which is human polynucleotide phosphorylase (hPNPase old-35 ), a 3Ј,5Ј exoribonuclease involved in RNA degradation (8). hPNPase old-35 is a highly evolutionary conserved gene in plants, prokaryotes and eukaryotes having similar domain structure and functional properties in all species. In vitro assays confirmed that hPNPase old-35 is involved in RNA degradation. Analysis of the expression profile of hPN-Pase old-35 revealed that it is predominantly a type I interferoninducible gene, and its expression is also induced in senescent fibroblasts in comparison with young fibroblasts. These findings indicate that hPNPase old-35 might play an essential role in interferon-and senescence-induced growth arrest. Indeed, when hPNPase old-35 is transfected by plasmid or transduced via a replication-incompetent adenovirus (Ad.hPNPase old- 35 ), there is a marked reduction in the colony-forming ability of HO-1 cells. The objective of the present study was to elucidate the molecular mechanism of the growth-suppressing property of Ad.hPNPase old- 35 .
The protooncogene c-myc is involved in a wide range of cellular processes including proliferation, differentiation, and tumorigenesis (for a recent review, see Ref. 9). Myc belongs to the Max network, a group of transcription factors containing basic helix-loop-helix zipper motifs (10,11). Myc heterodimerizes with Max and binds to the E-box sequence (CACGTG), thereby activating transcription (12). Max is constitutively expressed throughout the cell cycle (13), and it also heterodimerizes with the Mad family of transcription factors: Mad1, Mxi1, Mad3, and Mad4 (14 -16); however, in contrast to Myc-Max, the Mad-Max heterodimers act as transcriptional repressors at the same binding sites. The most important function of Myc is its essential role in controlling cell proliferation. Expression of exogenous Myc in cultured fibroblasts promotes S phase entry and shortens G 1 phase of the cell cycle, whereas activation of a conditional Myc is sufficient to drive quiescent cells into the cell cycle (17,18). The progression of cell cycle beyond the G 1 phase is also augmented by the activities of the cyclin-dependent kinase (CDK) complexes cyclin D-CDK4 and cyclin E-CDK2, and the activities of these complexes are inhibited by CDK inhibitors, the Cip/Kip (CDK-interacting protein/kinase-inhibitory protein) family, including p27 KIP1 and p21 CIP1/WAF-1/MDA-6 and the INK (inhibitors of CDK4) family including p16 INK4A and p15 INK4B (19). Cyclins D and E are essential for G 1 -S progression in higher eukaryotic cells and when overexpressed are able to shorten the G 1 interval (20,21). The major pathway by which Myc induces cell cycle progression is by activating cyclin D2 and CDK4 (22)(23)(24)(25). An important consequence of the induction by Myc of cyclin D2 is its sequestration of p27 KIP1 CDK inhibitor, permitting unfettered and prolonged activity of the cyclin E-CDK2 complex (26). Increased cyclin E-CDK2 activity shortens G 1 , whereas increased CDK2 and CDK4 activities result in hyperphosphorylation of the retinoblastoma (Rb) protein. This leads to release of E2F, a family of transcription factors that regulate a battery of genes necessary for cell cycle progression, from complexes with Rb and together with the direct induction of E2F2 by Myc, may further contribute to cell cycle progression (27). In addition Myc can directly repress p27 KIP1 and p21 CIP1/WAF-1/MDA-6 transcription (28,29).
In the present study, we show that infection of HO-1 melanoma cells with Ad.hPNPase old-35 resulted in cell cycle arrest in the G 1 phase and eventually apoptosis with marked reduction in DNA synthesis. Ad.hPNPase old-35 infection caused reduction in expression of the c-myc mRNA and Myc protein that was accompanied by induction of Mad1 protein. Overexpression of Myc protected HO-1 cells against Ad.hPNPase old-35 -mediated cell death. These findings argue that Myc might play a pivotal role in mediating the growth-inhibitory properties of Ad.hPN-Pase old-35 . We also show that inhibition of hPNPase old-35 by an antisense strategy partially but significantly rescues HO-1 cells from interferon-␤-mediated growth inhibition, thereby documenting a potential role of hPNPase old-35 in mediating interferon action.
Virus Construction and Infection Protocol-The construction and purification of hPNPase old-35 expressing replication-defective adenovirus Ad.hPNPase old-35 were described previously (8,31). A similar method was employed to generate an antisense hPNPase old-35 expressing replication-defective adenovirus (Ad.hPNPase old-35 AS). The empty adenoviral vector (Ad.vec) was used as a control. Viral infections were performed as previously described (30).
Plasmid Construction, Transfection, and Colony Formation Assays-3Ј-HA-tagged hPNPase old-35 was created by PCR using the primers GCT AGC ATG GCG GCC TGC AGG TAC (sense) and GGA TCC TCA AGC GTA ATC TGG AAC ATC GTA TGG GTA CTG AGA ATT AGA TGA TGA (antisense). The authenticity of the amplified product was verified by sequencing, and it was cloned into the NheI/BamHI sites of pcDNA3.1 (Invitrogen) to generate hPNPase old-35 -HA. hPNPase old-35 AS was generated by ligating hPNPase old-35 in an antisense orientation into BamHI/NotI sites of pREP4 (Invitrogen). The c-myc expression plasmid p290-myc (2, 3) was provided by Dr. Riccardo Dalla-Favera. HO-1 cells were plated at a density of 3 ϫ 10 5 cells/6-cm dish and 24 h later were transfected with 5 g of either empty vector or p290-myc (2, 3) using Superfect ® (Qiagen, Hilden, Germany) transfection reagent according to the protocol from the manufacturer. After 36 h, the cells were infected with Ad.hPNPase old-35 at a multiplicity of infection (m.o.i.) of 50 or 100 pfu/cell; 6 h later, the cells were trypsinized and counted and 10 3 cells were plated in 6-cm dishes. Colonies were counted after 3 weeks. Colony formation assays using hPNPase old-35 -HA in HO-1 cells were performed as described (32).
RNA Isolation and Northern Blot Analysis-Total RNA was extracted from the cells using Qiagen RNeasy mini kit (Qiagen) according to the protocol from the manufacturer, and Northern blotting was performed as described (33). The cDNA probes used were a 400-bp fragment from human c-myc, a 500-bp fragment from hPNPase old-35 , a 500-bp fragment from human GADD34, full-length human c-jun, and full-length human GAPDH.
In Vitro Translation and in Vitro mRNA Degradation Assays-In vitro translation was performed using the TNT-coupled Reticulocyte Lysate Systems (Promega, Madison, WI) using the plasmids pcDNA3.1 as a control, GADD153 expression plasmid, and hPNPase old-35 -HA according to the protocol from the manufacturer. Five g of total RNA from HO-1 cells were incubated with 5 l of each in vitro translated protein at 37°C from 0.5 to 3 h. The RNA was repurified using the Qiagen RNeasy mini kit, and Northern blotting was performed (33).
Western Blot Analysis-Western blotting was performed as previously described (33). Briefly, cells were harvested in radioimmune precipitation assay buffer containing protease inhibitor mixture (Roche Molecular Biochemicals, Mannheim, Germany), 1 mM Na 3 VO 4 , and 50 mM NaF and centrifuged at 12,000 rpm for 10 min at 4°C. The supernatant was used as total cell lysate. Thirty g of total cell lysate were used for SDS-PAGE and transferred to a nitrocellulose membrane. Cell Cycle Analysis-Cells were harvested, washed in phosphatebuffered saline , and fixed overnight at Ϫ20°C in 70% ethanol. The cells were treated with RNase A (1 mg/ml) at 37°C for 30 min and then with propidium iodide (50 g/ml). Cell cycle was analyzed using a FACScan flow cytometer, and data were analyzed using CellQuest software (Becton Dickinson, San Jose, CA).
Telomerase Assay-HO-1 cells were infected with either Ad.vec or Ad.hPNPase old-35 for 1-4 days or untreated or treated with fibroblast IFN-␤ (2000 units/ml) plus MEZ (10 ng/ml) for 1-4 days, and telomerase assays were performed as described previously (34). Briefly, protein concentrations of cell extracts were determined, and equal amounts of protein were used for the elongation process in which telomerase added telomeric repeats (TTAGGG) to the 3Ј end of the biotin-labeled primer. These elongation products were amplified by PCR, and the PCR products were denatured and hybridized to digoxigenin-labeled detection probes, specific for the telomeric repeats. The resulting products were immobilized via the biotin label to a streptavidin-coated microtiter plate. Immobilized amplicons were detected with an antibody against digoxigenin that is conjugated to horseradish peroxidase and the sensitive peroxidase substrate 3,3Ј,5,5Ј-tetramethylbenzidine. The telomerase activity was quantified by measuring the absorbance of the samples at 450 nm (with a reference wavelength of 690 nm) using a microtiter plate reader.
Statistical Analysis-Statistical analysis was performed using oneway analysis of variance, followed by Fisher's protected least significant difference analysis.

RESULTS
Previous studies demonstrated that infection with Ad.hPN-Pase old-35 inhibited colony formation in HO-1 melanoma cells (8). The present studies were conducted to comprehend the molecular mechanism underlying the growth-arresting property of Ad.hPNPase old- 35 . Different melanoma cell lines and SV40 T-Ag immortalized primary human melanocytes (FM-516-SV) were infected with Ad.hPNPase old-35 , and the growth of the cells was monitored by standard MTT assays. As shown in Fig. 1, infection with Ad.hPNPase old-35 resulted in significant growth inhibition in all of the cells. The growth-inhibitory effect became significant from 4 days after infection, and, in certain cell lines (WM278 and MeWo), infection with Ad.hPN-Pase old-35 completely inhibited cell growth. In addition, Ad.hPN-Pase old-35 infection inhibited the growth of other cell types, including breast, prostate, colon and pancreatic carcinomas, glioblastoma multiforme, fibrosarcoma, and osteosarcoma, irrespective of their p53 or Rb status (data not shown).
To investigate the mechanism of Ad.hPNPase old-35 -mediated growth inhibition, cell cycle analysis was performed following Ad.hPNPase old-35 infection in HO-1 cells. When the cells were infected with Ad.hPNPase old-35 at a high m.o.i. of 100 pfu/cell for 4 days, there was a significant increase in sub-G 0 population of cells indicating apoptosis and a decrease in the S phase indicating inhibition of DNA synthesis (Fig. 2, A and B). When the kinetics of killing was slowed down by infecting cells at a low m.o.i. of 25 pfu/cell, there was an initial significant increase in cells in the G 1 phase of the cell cycle (Fig. 2, C and D) at 3 and 6 days after infection. This increase was also accompanied by a marked decrease in the S phase. At later time points, the cells infected with Ad.hPNPase old-35 , but not control or Ad.vecinfected cells, started to die by apoptosis (Fig. 2D). The kinetics of cell death was very slow when HO-1 cells were infected with Ad.hPNPase old-35 at a low m.o.i. It is worth noting that at 25 pfu/cell ϳ90% of the cells are infected with adenovirus (data not shown). From these observations it might be inferred that infection with Ad.hPNPase old-35 induces cell cycle arrest at the G 1 phase that ultimately culminates in apoptosis.
The inhibition of DNA synthesis following Ad.hPNPase old-35 infection was confirmed using a [ 3 H]thymidine incorporation assay. As shown in Fig. 3A, infection with Ad.vec did not have an impact on DNA synthesis. Infection with Ad.hPNPase old-35 reduced DNA synthesis by ϳ40% at an m.o.i. of 25 pfu/cell and by ϳ75% at 50 pfu/cell 4 days after infection.
Telomerase activity is decreased in both terminal differentiation and senescence. As shown in Fig. 3B, telomerase activity decreased in a time-dependent manner to ϳ50% when HO-1 cells were treated with IFN-␤ ϩ MEZ for up to 4 days. This treatment protocol results in the induction of irreversible growth arrest and terminal differentiation in HO-1 melanoma cells (2,5). Based on these findings, telomerase activity was also determined following Ad.hPNPase old-35 infection. As shown in Fig. 3C, infection with Ad.hPNPase old-35 at an m.o.i. of 100 pfu/cell, but not with Ad.vec, inhibited telomerase activity by almost 60% at day 4 after infection.
One of the factors that facilitate entry into the S phase of the cell cycle is Myc. During terminal differentiation of melanoma cells, c-myc mRNA expression is down-regulated (35). The expression level of c-myc mRNA following Ad.hPNPase old-35 infection was, therefore, determined by Northern blot analysis. The expression of c-myc mRNA began decreasing 2 days after Ad.hPNPase old-35 infection but not in uninfected or Ad.vecinfected cells even at 4 days after infection (Fig. 4A). This decrease correlates with the expression of hPNPase old-35 mRNA that was also detected 2 days after infection. It should be noted that, under basal condition, hPNPase old-35 mRNA is undetectable in HO-1 cells. The expression of the GAPDH housekeeping gene remained unchanged following Ad.hPN-Pase old-35 infection.
Down-regulation of Myc protein by different stimuli is usually accompanied by up-regulation of Mad1, the transcriptional repressor belonging to the Max family of transcription factors (9). In this context, the expressions of Myc, its heterodimer partner Max, and Mad1 were determined by Western blot analysis following Ad.hPNPase old-35 infection. As anticipated from Northern blot analysis, Myc expression started decreasing 2 days after Ad.hPNPase old-35 infection but not in uninfected or Ad.vec-infected cells at 4 days after infection (Fig. 4B). hPNPase old-35 is a 3Ј,5Ј-exoribonuclease, prompting us to determine whether it can directly degrade c-myc mRNA. For this analysis a C-terminal HA-tagged hPNPase old-35 -expressing construct (hPNPase old-35 -HA) was created. The authenticity of the construct was first confirmed by transfecting it into HEK-293 cells. As shown in Fig. 5A, Western blot analysis using anti-HA antibody detected a single protein of ϳ90 kDa in size only in the hPNPase old-35 -HA-transfected cells. To check whether this construct has functional similarity to Ad.hPN-Pase old-35 , HO-1 cells were transfected with hPNPase old-35 -HA and cell growth was monitored by colony formation assay. As shown in Fig. 5B, overexpression of hPNPase old-35 -HA reduced colony-forming ability by ϳ45% indicating that hPN-Pase old-35 -HA also has growth-suppressing properties. After establishing that hPNPase old-35 -HA generates functional protein, this construct was used to prepare in vitro translated hPNPase  . The effect of hPNPase OLD-35 protein on c-myc mRNA was investigated by in vitro mRNA degradation assays. As shown in Fig. 5C, incubation with in vitro translated hPN-Pase OLD-35 resulted in degradation of c-myc mRNA. This effect was specific for c-myc because the mRNAs for the GADPH housekeeping gene, cell growth regulatory gene c-jun, and apoptosis-inducing gene GADD34 were not degraded. This effect was also specific for hPNPase OLD-35 because incubation with in vitro translated GADD153, a transcription factor, did not result in mRNA degradation.  (2,5). Infection with Ad.hPNPase old-35 AS provided small but significant protection against IFN-␤-induced growth inhibition both in HO-1-pREP4 and in HO-1-hPNPase old-35 AS in a dose-dependent manner (Fig. 7, B and D). As shown in Fig. 7B, treatment with IFN-␤ reduced cell viability by ϳ75% in HO-1-pREP4. In HO-1-pREP4 infected with Ad.hPNPase old-35 AS at an m.o.i. of 100 pfu/cell and in HO-1-hPNPase old-35 AS, cell viability was reduced by ϳ55% by IFN-␤ (Fig. 7B, columns 4 and 5, respectively). When HO-1-hPNPase old-35 AS cells were infected with Ad.hPNPase old-35 AS at an m.o.i. of 100 pfu/cell, the cell viability was reduced by only ϳ45% (Fig. 7B, column 8). A combination of IFN-␤ and MEZ treatment reduced cell viability of HO-1-pREP4 by ϳ94% (Fig. 7D, column 1). This inhibitory effect was partially reversed in HO-1-pREP4 cells infected with Ad.hPNPase old-35 AS at an m.o.i. of 100 pfu/cell, and in HO-1-hPNPase old-35 AS cells, the reduction in cell viability was ϳ87% (Fig. 7D, columns 4 and 5, respectively). In HO-1-hPN-Pase old-35 AS cells infected with Ad.hPNPase old-35 AS at an m.o.i. of 100 pfu/cell, the reduction in cell viability was ϳ80% (Fig. 7D, column 8). The antisense approach did not provide protection against MEZ-induced growth arrest (Fig. 7C) (Fig. 8A). The expression levels of the anti-apoptotic protein Bcl-2 and pro-apoptotic protein Bax remained unchanged. Stable HO-1 cell lines expressing either Bcl-2 or Bcl-xL were generated, and these cell lines were infected with Ad.hPNPase old-35 . As shown in Fig. 8B, overexpression of Bcl-xL, but not Bcl-2, provided partial protection against Ad.hPNPase old-35 -induced cell death.

DISCUSSION
The molecular mechanism by which Ad.hPNPase old-35 induces growth arrest has striking resemblance to that observed during terminal differentiation and senescence. The Max network of transcription factors has been implicated in mediating cell cycle arrest during differentiation. Myc levels rapidly diminish during terminal differentiation of many cell types, and enforced expression of Myc inhibits or modulates terminal differentiation of myoblasts, erythroleukemia cells, adipocytes, B lymphoid cells, and myeloid cells among others (36 -39). Inhibition of Myc is able to block mitogenic signals and drive cells toward terminal differentiation. Stable transfection of multiple copies of the Myc DNA-binding sequence, resulting in sequestration and squelching of endogenous Myc, accelerates proliferative senescence in rat embryo fibroblasts (40). During the transition from an undifferentiated to a differentiated state, there is a switch from Myc-Max to Mad-Max complexes that function as transcriptional repressors. Overexpression of the Mad family transcription factors Mad1 and Mxi1 arrests cells in the G 1 phase of the cell cycle and promotes differentiation of some cell types (41)(42)(43). There is an intimate relationship between members of the Max network and the CDK inhibitors.
The expression of both Mad1 and p27 KIP1 are up-regulated during differentiation (44 -47). Terminally differentiated melanocytes induced by cAMP display elevated levels of p27 KIP1 (48). Moreover, Myc is down-regulated during differentiation as levels of p27 KIP1 increase. It has been demonstrated that Mad1 and p27 KIP1 cooperate to promote terminal differentiation of granulocytes and to inhibit Myc expression and cyclin E-CDK2 activity (49). Overexpression of Mad1 in human melanoma cells results in a reduced proliferation rate, increased G 0 /G 1 cell cycle accumulation, and active melanin synthesis (50). Our observation that Ad.hPNPase old-35 infection results in downregulation of Myc and up-regulation of Mad1 and that overexpression of Myc can protect from Ad.hPNPase old-35 -induced cell death suggests the involvement of the Max network in this process. More importantly, we observed that telomerase activity decreases following Ad.hPNPase old-35 infection, which is also observed during terminal differentiation and senescence (51). During the induction of terminal differentiation in HO-1 cells by IFN-␤ ϩ MEZ treatment, telomerase activity was also decreased (Fig. 3B). It has been shown that human telomerase catalytic subunit (hTERT) is transcriptionally regulated by Myc-Max (52). Thus, down-regulation of Myc can partially explain the decreased telomerase activity following Ad.hPNPase old-35 infection.
There are several models of growth arrest in which downregulation of Myc is associated with up-regulation of p27 KIP1 . Retinoic acid-induced G 1 arrest of human myeloid cells results in sequential down-regulation of Myc and cyclin E and upregulation of p27 KIP1 (53). When deprived of adhesion, mam-mary epithelial cells arrest in the G 1 phase of the cell cycle (54). This arrest is associated with down-regulation of Myc and up-regulation of p27 KIP1 and overexpression of Myc in nonadherent cells reverses the inhibition of cell cycle progression. One interesting finding of the present study is that, although the level of p27 KIP1 increased, that of p21 CIP1/WAF-1/MDA-6 decreased in response to Ad.hPNPase old-35 infection in HO-1 cells. Although both of these CDK inhibitors are involved in G 1 arrest, the association of Myc with p27 KIP1 is more profound than that with p21 CIP1/WAF-1/MDA-6 . In addition to being a transcriptional repressor for p27 KIP1 , Myc also drives the synthesis of a putative p27 KIP1 -sequestering protein, which renders the hyperphosphorylated p27 KIP1 available for ubiquitination (55). Accordingly, it is likely that the down-regulation of Myc protein could prevent the synthesis of this putative protein, making p27 KIP1 unavailable for ubiquitination. p27 KIP1 is a recognized target for SCF, the ubiquitin ligase, and one of the critical components of SCF is Cul1, a Myc target gene (56). Thus, down-regulation of Myc results in increased p27 KIP1 by multiple pathways. Two models of senescence-like growth arrest, one because of iron chelation in hepatocytes, the other because of inhibition of phosphoinositide 3-kinase pathway in mouse embryo fibroblasts, are also associated with up-regulation of p27 KIP1 and down-regulation of p21 CIP1/WAF-1/MDA-6 (57, 58). In both cases, the levels of p53 and p16 INK4A were also decreased. We could not detect p16 INK4A in HO-1 cells because the major- ity of melanomas are associated with genetic abnormalities in the p16 INK4A locus. By Western blot analysis we could not detect any change in p53 protein level following Ad.hPN-Pase old-35 infection. Thus, we could not attribute the downregulation of p21 CIP1/WAF-1/MDA-6 , which lies downstream of p53, as a result of down-regulation of p53. Interestingly, we observed that both p21 CIP1/WAF-1/MDA-6 and p53 were downregulated at the mRNA level after Ad.hPNPase old-35 infection (data not shown). A detailed study monitoring the half-life of p53 mRNA and protein, its phosphorylation status, and transcription-inducing activity following Ad.hPNPase old-35 infection might be required to address the issue of whether there is any association between p53 and subsequent p21 CIP1/WAF-1/MDA-6 down-regulation.
In the present study we observed that Ad.hPNPase old-35 infection induces initial cell cycle arrest but eventually the cells die by apoptosis. Inhibition of c-myc in M14 human melanoma cells by an inducible antisense approach resulted in G 1 cell cycle arrest, which was associated with increased levels of p27 KIP1 and decreased levels of p21 CIP1/WAF-1/MDA-6 (59). These cells eventually became apoptotic, and it was shown that upregulation of p27 KIP1 was directly responsible for promoting apoptosis. It has been shown that, in addition to inhibiting cell cycle progression, overexpression of p27 KIP1 alone could induce apoptosis (60,61). Overexpression of Bcl-2 can protect HeLa cells from p27 KIP1 -induced apoptosis (61). Ad.hPNPase old-35 infection did not alter the level of Bcl-2 protein in HO-1 cells; however, there was a significant reduction in Bcl-xL protein level and overexpression of Bcl-xL in these cells could protect HO-1 cells from Ad.hPNPase old-35 -mediated cell death. Bcl-2 and Bcl-xL are functionally similar molecules, and they are often functionally interchangeable, depending upon the cell type. In fact, in HO-1 melanoma cells, it has been shown previously that apoptosis-inducing agents modulate the level of Bcl-xL protein at a higher level than that of Bcl-2 protein (30). 3Ј,5Ј processing or degradation of RNA controls many cellular events including the maturation of 5.8 S rRNA, the processing of many small RNAs, and the turnover of different types of mRNAs (62)(63)(64)(65). In eubacteria and chloroplasts, 3Ј,5Ј RNA decay occurs through a multiprotein complex called the degradosome (66,67). The bacterial degradosome contains several proteins, including the endoribonuclease RNase E, the exoribonuclease PNPase, enolase, and an RNA helicase (RhlB) (68,69). The chloroplast degradosome contains only PNPase and no other proteins (67). In eukaryotic cells this processing is performed by the exosome which consists of at least 11 proteins, all of which either possess 3Ј,5Ј-exoribonuclease activity or are predicted to contain exoribonuclease or RNA binding activity based on their homology with prokaryotic proteins such as RNase PH, RNase R, RNase D, and the RNA-binding domain of the S1 ribosomal protein (63,70). The yeast exosome does not contain PNPase; however, the structural model based on the protein-protein interactions of the exosome subunits reveal that it resembles the structure of bacterial PNPase, which form homotrimers (71). With the isolation of the human exosome components, a similar structural model for human exosome has also been proposed (72). Although this model excludes the involvement of hPNPase old-35 in the human exosome, it might be possible that hPNPase old-35 has evolved to regulate a more specific function as is documented by its involvement in growth arrest during terminal differentiation and senescence. Another possible specialized function of hPNPase old-35 could be its involvement in an alternative IFN-stimulated RNA-decay pathway that might include hPNPase old-35 , mda-5, a putative RNA helicase having growth suppressive property and mda-E-63, a gene having an RNase II motif (32,73). Our observation that antisense inhibition of hPNPase old-35 rescues the cells from IFN-␤-mediated growth inhibition confirms its biological role in mediating interferon action.
Terminal differentiation and senescence are two of the important ways by which the human body protects itself from the lethal effects of carcinogenesis. The observation that hPN-Pase old-35 lies at the crossroads of these two important physiological processes and induces growth arrest supports a potential tumor suppressor role for this molecule. Further rigorous in vivo studies are required to confirm this possibility and to establish the functional properties of hPNPase old- 35 . These studies will also indicate whether this gene and its encoded protein can be used for therapeutic applications. Experiments in fibroblasts and primary human melanocytes are also necessary to address the question of whether overexpression of hPN-Pase old-35 can induce a true senescence-like phenotype and also whether specific differentiation markers are altered during this process. These studies should be very enlightening and may provide new insights into the processes of terminal differentiation and senescence and the functional relationships between these alternative cellular fates.