Introduction

Members of the transforming growth factor- (TGF-) superfamily potently inhibit the proliferation of many types of cells (Attisano and Wrana, 2000; Massague et al., 2000; Wrana and Attisano, 2000). TGF- binds the cell-surface types I and II receptors, both of which are serine/threonine protein kinases. Upon binding TGF-, the type II receptor phosphorylates the type I receptor. The Smad anchor for receptor activation (SARA), a membrane-associating protein, escorts unphosphorylated Smad2 and Smad3 to the receptor (Tsukazaki et al., 1998), which in turn phosphorylates these Smad proteins. On the other hand, caveolin-1, a protein associating with caveolae, downregulates TGF- signaling by interacting with the type I receptor (Razani et al., 2001). The TGF- type III receptor also facilitates TGF- signaling, which is regulated by a PDZ domain-containing protein (Wang et al., 1991; Blobe et al., 2001). The phosphorylated Smad2 and Smad3 proteins associate with Smad4 and enter the nucleus; together with other transcription factors, they activate transcription of various target genes (Lagna et al., 1996; Zhang et al., 1996; Liu et al., 1997b; Derynck et al., 1998), leading to cell cycle arrest at the G1 phase.

Since Smad3 and Smad4 bind to only four base pairs of DNA sequence, AGAC or GTCT (Shi et al., 1998; Zawel et al., 1998), they usually cooperate with other transcription factors to increase their specificity in the regulation of gene expression (Derynck et al., 1998; Massague et al., 2000). For example, activated Smad2 and Smad4 interact with FAST1 to turn on the transcription of a developmentally regulated gene, Mix.2 (Chen et al., 1996). Smad3 and Smad4 also interact with the AP-1 complex to activate gene transcription (Zhang et al., 1998; Liberati et al., 1999). They associate with TFE3, a basic helix–loop–helix transcription factor, and Sp1, a zinc finger-containing transcription factor, to activate transcription of the gene for plasminogen activator inhibitor type 1 (Hua et al., 1998; Datta et al., 2000). Activated Smad2 and Smad3 interact with Sp1 to activate transcription of cyclin-dependent kinase inhibitors including p15^ink4b (Hannon and Beach, 1994; Li et al., 1995; Moustakas and Kardassis, 1998; Feng et al., 2000). Many other transcription factors, such as Evi-1, vitamin D₃ receptor, androgen receptor, BF-1, and c-Myc, have been shown to interact with Smad proteins to regulate the Smads' activity and enhance their specificity (Kurokawa et al., 1998; Yanagisawa et al., 1999; Hayes et al., 2001; Chipuk et al., 2002; Feng et al., 2002; Ten Dijke et al., 2002). Moreover, various transcriptional coactivators and corepressors, including p300 and Ski/Sno, interact with Smad proteins to further regulate the activity of Smad proteins (Feng et al., 1998; Janknecht et al., 1998; Pouponnot et al., 1998; Shen et al., 1998; Luo et al., 1999; Sun et al., 1999; Kim et al., 2000).

A critical role for TGF- in the suppression of tumorigenesis has been well documented. In colon cancer with microsatellite instability, mutations in the type II receptor gene are frequently detected (Markowitz et al., 1995). Mutations in the Smad2 and Smad4 genes have also been identified in human colorectal cancer cells (Eppert et al., 1996). The Smad4 (DPC4) gene is frequently mutated in pancreatic cancer (Hahn et al., 1996). Mice with targeted disruption of the Smad3 gene develop multiple colon cancers (Zhu et al., 1998; Datto et al., 1999; Yang et al., 1999). Moreover, compound mutations in the Smad4 gene and the tumor suppressor gene APC also lead to colon cancer in mice (Takaku et al., 1998). Recently, the Smad4 gene has been reported to be mutated in human acute promyelocytic leukemia (APL) cells (Imai et al., 2001). AML1/ETO, a chimeric transcription factor involved in acute myelogenous leukemia, was shown to interact with Smad3 and to attenuate TGF--induced transcription (Jakubowiak et al., 2000). Together, these results clearly suggest a critical role for the TGF- pathway in suppressing tumorigenesis.

Retinoic acid receptor (RAR) is a transcription factor that is activated by binding to retinoic acid (RA). RAR usually forms a dimer with retinoid X receptor (RXR), and the resulting complex binds to the promoter of its target genes, leading to their transcription in the presence of RA (Mangelsdorf and Evans, 1995). Translocations of the RAR gene, which usually involve the formation of a fusion protein between RAR and other proteins due to chromosomal translocations, are usually associated with APL (Look, 1997; Slack, 1999; Pandolfi, 2001). PML–RAR and PLZF–RAR are two commonly identified fusion proteins in acute promyelocytic leukemia (APL) (Look, 1997; Pandolfi, 2001). Although oncogenic PLZF–RAR is able to bind RA as well as the DNA sequence that is normally recognized by RAR, it actively inhibits the transcription of the target genes of the normal RAR. The PLZF portion of the PLZF–RAR fusion protein has been shown to interact with transcriptional corepressors including N-CoR/SMRT, mSin3, and histone deacetylase (Hong et al., 1997; Grignani et al., 1998; Guidez et al., 1998; He et al., 1998; Lin et al., 1998). RA also inhibits proliferation of cervical cancer cells and some lung cancer cells (Meyskens and Manetta, 1995), while the RAR gene is often silenced by methylation of its promoter DNA in RA-resistant lung cancer cell lines (Geradts et al., 1993; Virmani et al., 2000).

Cooperation between TGF- and RA has been reported to occur in both physiological and pathological contexts (Roberts and Sporn, 1992; Dickens and Colletta, 1993). For example, TGF- and RA together inhibit growth of retinal pigment epithelial cells (Kishi et al., 2001), bone marrow progenitor cells (Lardon et al., 1996), and human breast cancer cells (Valette and Botanch, 1990). On the other hand, they can also cooperate to induce fibrosis from hepatic stellate cells by enhancing the synthesis of procollagen (Okuno et al., 1997,2002). RA induces activation of latent TGF- and enhances expression of TGF- and its receptor (Glick et al., 1989; Falk et al., 1991; Batova et al., 1992; Morales and Roberts, 1992; Kojima and Rifkin, 1993; Danielpour, 1996; Han et al., 1997; Imai et al., 1997). Thus, a prevalent model for the relationship between RA and TGF- is that RA increases production of active TGF- and its receptors, leading to growth inhibition. Supporting this hypothesis, neutralizing antibodies against TGF- can block RA-induced growth inhibition of several types of cells, including keratinocytes and human U937 leukemic cells (Nunes et al., 1996; Defacque et al., 1999; Borger et al., 2000). However, it is unclear whether there are any intracellular interactions between TGF- and RA pathways.

Through a functional cloning strategy using a retroviral cDNA library, we found that the fusion between gag and RXR was responsible for resistance to TGF--induced growth inhibition in mink lung epithelial cells. We also show that PLZF–RAR fusion protein mimics the properties of gag–RXR and thus antagonizes TGF--induced growth inhibition. Moreover, TGF- and RA cooperate at an intracellular level to inhibit growth of mink lung epithelial cells as well as induce transcription of p15^INK4b, while expression of PLZF–RAR attenuates TGF--induced expression of p15^INK4b. Together, these findings suggest that TGF- and RA normally cooperate to inhibit cell proliferation and that expression of PLZF–RAR leads to resistance to TGF--induced growth inhibition.

Results

Expression cloning of RXR from cells that are resistant to TGF--induced growth inhibition

To identify novel components in the TGF- signal transduction pathway, we utilized a retrovirus-mediated expression cloning strategy to isolate cDNAs whose overexpression leads to resistance to TGF--induced growth inhibition. To this end, we made a retroviral cDNA library with mRNA from mink lung epithelial cells and infected 10⁷ ML-ER5 cells, the MvlLu cells expressing a murine retrovirus ectopic receptor, with a retroviral cDNA library, and then subjected the infected cells to TGF- selection. We isolated 10 cell clones that were resistant to TGF--induced growth inhibition. From the genomic DNA of the first TGF--resistant clone, the retrovirus-transduced cDNA was identified by polymerase chain reaction (PCR) amplification, which was the cDNA for the TGF- type II receptor. Further sequence analysis revealed that the type II receptor cDNA stops after the transmembrane domain (at amino acid 261) and thus lacks the essential serine/threonine kinase domain (data not shown), a well-known dominant negative form of the receptor (Bottinger et al., 1997). These initial results suggested that the expression cloning approach was able to isolate components that actively participate in TGF- signal transduction.

RXR, a partner for RAR, was isolated from two independent TGF--resistant cell clones. Figure 1a shows that these two cell clones, clones 5 and 19, were independently infected by recombinant retroviruses carrying the RXR cDNA, since cell clone 5, but not clone 19, harbors an additional retrovirus-transduced cDNA (Figure 1a). The coding sequence of the mink RXR (m-RXR) cDNA was sequenced, and the results show that mink RXR has high homology to human and mouse RXR, 98.7 and 97.5%, respectively (data not shown). The DNA sequences from cell clones 5 and 19 are identical, with both lacking the first seven amino-acid residues.

Figure 1.

Cloning of RXR from TGF--resistant mink lung epithelial cell clones. (a) PCR amplification of mink RXR from two independent TGF--resistant cell clones. Mink RXR cDNA was amplified from the genomic DNA from the resistant cell clones 5 and 19, using a pair of primers flanking the multiple cloning site and the retroviral vector pMX, which are XH15 and MX3. The asterisk (*) denotes a PCR-amplified fragment for an RNA spliceosome factor that was transduced by another retrovirus. The accession number for the mink RXR is AF418003. (b) RXR in a retroviral vector causes resistance to TGF--mediated growth inhibition. ML-ER5 cells infected with control viruses (vector) or viruses carrying pMX-RXR were seeded in a 100 mm dish at a density of 10⁶ cells/dish on day 0. On day 1, TGF- was added to each dish at a concentration of 10 pM. Medium was changed every 4 days. The cells were stained on day 12 with crystal violet. Living cells were stained dark violet

Full figure and legend (121K)

Next, we sought to confirm that the PCR-amplified RXR, when inserted into a retroviral vector, was indeed able to cause resistance to TGF--induced growth inhibition. To this end, ML-ER5 cells were infected with either control viruses or viruses carrying the RXR cDNA (from clone 19), and then subjected to TGF- selection, followed by crystal violet staining for viable cells. Figure 1b shows that ML-ER5 cells infected with control viruses did not survive the TGF- selection, but cells infected with viruses carrying the RXR cDNA were resistant to TGF-.

Fusion of retroviral gag and RXR is essential for resistance to TGF--induced growth inhibition

Since RXR dimerizes with RAR to regulate gene transcription and inhibit cell proliferation (Dragnev et al., 2001), it is paradoxical that the expression of RXR causes resistance to TGF-. Analysis of the cloned RXR DNA sequence revealed that the eighth codon of the mink RXR was fused in frame to the retroviral gag sequence that is important for packaging of retroviral particles. Thus, we suspected that the recombinant retrovirus carrying the RXR sequence generated a gag–RXR fusion protein in cells and that this fusion protein might act as a dominant negative mutant, causing resistance to TGF--mediated growth inhibition in cells. To test this hypothesis, we epitope-tagged the NH₂-terminus of the retroviral gag sequence in the retroviral vector and the NH₂-terminus of the RXR sequence (Figure 2a, left panel). Subsequently, the two constructs, Myc-gag–RXR and Myc-RXR, were transfected into human 293 cell-derived E-NX cells, followed by immunoblotting analysis with an anti-Myc antibody. Figure 2a (right panel) shows that indeed a fusion protein of myc-gag–RXR with an expected size of 87 kDa (lane 3) was expressed in the transfected cells, as compared to that of myc-RXR (56 kDa, lane 2).

Figure 2.

Gag–RXR fusion protein, but neither gag nor RXR alone, causes resistance to TGF--mediated inhibition of cell growth. (a) Detection of the gag–RXR fusion protein. Left panel, a diagram of the myc epitope-tagged gag–RXR and RXR constructs. Right panel, Western blot analysis of myc-gag–RXR and myc-RXR proteins from E-NX cells. On day 0, 810⁵ E-NX cells were seeded in each well of six-well plates. On day 1, the cells were transfected with 2.5 g of myc-gag–RXR, myc-RXR or vector (pMX-Puro) as indicated. On day 2, transfected cells were switched to fresh medium. On day 4, the cells were harvested for the preparation of nuclear extracts. For immunoblotting analysis, 30 g protein from the nuclear extract was loaded in each lane for separation, and blotted with the antibody against c-myc epitope (9E10). The asterisk (*) denotes a polypeptide that is also recognized by the 9E10 monoclonal antibody. (b) A diagram of gag–RXR and its derivative constructs with various insertions between gag and RXR. Gag–RXR was expressed from pMX-RXR. (c) Western blot analysis of gag–RXR and its derivatives in E-NX cells transfected with the indicated constructs. The total amount of proteins from the nuclear extracts loaded in each lane was 75 g. The membrane was blotted with a polyclonal antibody against mouse RXR (Upstate Biotechnology, #06-527). (d) ML-ER5 cells were infected with retroviruses carrying gag–RXR or its derivative sequences as indicated in (b). The infected cells were first selected with 2 g/ml puromycin, and puromycin-resistant cells were seeded at a density of 510⁵ cells per 100 mm dish and further treated with 30 pM TGF-. After 12 days of TGF- selection, the cells were stained with crystal violet

Full figure and legend (153K)

To test whether the fusion of gag and RXR or the lack of the first seven amino-acid residues of the mink RXR causes resistance to TGF--induced growth inhibition, we generated the following various constructs that should affect the formation of the gag–RXR fusion protein as indicated in Figure 2b: (1) insertion of one nucleotide, two nucleotides, or three nucleotides between the gag and the RXR sequences in a retroviral vector, (2) insertion of a three-nucleotide stop codon, (3) fusion of an initial ATG codon to the eighth amino-acid residue of the M-RXR (Figure 2b). The prediction was that all these insertions, except for TAC, would abrogate formation of the gag–RXR fusion protein. In contrast, insertion of the three nucleotides TAC, a codon for tyrosine, would generate a gag–RXR fusion protein with one additional amino-acid residue. The above constructs, as shown in Figure 2b, were individually transfected into E-NX cells and the expression of gag–RXR in transfected cells was analysed by immunoblotting analysis using an anti-RXR antibody. Figure 2c shows that, as expected, only the constructs for gag–RXR and gag+3–RXR generated the gag–RXR fusion protein (lanes 2 and 5), and all of the other constructs with insertions failed to generate any fusion protein (lanes 3, 4, and 6). M-RXR was also expressed at a high level but its size overlapped with the endogenous RXR (lane 7).

The resulting recombinant retroviruses, as indicated in Figure 2b, were used to infect ML-ER5 cells, and the infected cells were subjected to selection by TGF-, followed by crystal violet staining for living cells. Figure 2d clearly shows that only gag–RXR and gag+3–RXR, but not the other constructs that fail to generate gag–RXR fusion protein, cause resistance to TGF--induced growth inhibition. These results strongly suggest that it is the gag–RXR fusion protein, but neither M-RXR nor gag alone, that causes resistance to TGF--induced growth inhibition.

To quantitatively confirm resistance to TGF- conferred by gag–RXR, an individual cell subclone expressing gag–RXR and the pooled gag–RXR-expressing cells were treated with various concentrations of TGF- before counting cell numbers. Figure 3a shows that TGF- at a concentration of 20 pM potently inhibits growth of control cells by 90%; in contrast, TGF- at this concentration only inhibits the cell clone expressing gag–RXR by 10% and the pooled gag–RXR-expressing cells by 33%.

Figure 3.

Mink lung epithelial cells expressing gag–RXR become resistant to TGF- as well as RA-induced growth inhibition. Control cells and a representative cell clone or pooled clones expressing gag–RXR were plated on six-well plates (510⁴ cells/well) on day 0. On day 1, TGF- at the indicated concentrations was added to each well, and the cells were further cultured for 5 days before harvesting for counting with a Coulter automatic counter. This is a representative of three independent experiments. (b) Resistance of ML-ER5 cells expressing gag–RXR to RA. A pool of ML-ER5 cells infected with control retroviruses (control cells) and a pool of ML-ER5 cells infected with gag–RXR viruses were seeded in six-well plates at 510⁴ cells/well on day 0. On day 1, RA was added to each well at the indicated concentrations. On day 3, freshly made RA was added to corresponding wells. Cells were harvested on day 5 for counting. This is a representative of three independent experiments. (c) The cells were set up in 96-well plates (5000 cells/well) on day 0, and a medium with various concentrations of FBS was added to the cells to achieve indicated concentrations of FBS on day 1. On day 3, cells were pulsed by addition of 1 Ci (20 l)/well of ³H-thymidine for 4 h to determine the incorporation of ³H-thymidine into cells, as described in Materials and methods. This is a representative of three independent experiments

Full figure and legend (60K)

Gag–RXR confers resistance to RA-induced growth inhibition but does not enhance proliferation of the cells

Since RXR is a partner for RAR, we determined if cells expressing gag–RXR are also resistant to RA-induced inhibition of cell growth. Cells infected with either control viruses or viruses carrying gag–RXR were treated with increasing concentrations of RA (Figure 3b) and were counted after 4 days of culture. The results show that RA suppresses growth of normal cells but fails to inhibit growth of cells expressing gag–RXR (Figure 3b). These results indicate that gag–RXR not only blocks TGF--induced growth inhibition but also causes resistance to RA-mediated growth inhibition. Control cells and cells expressing gag–RXR proliferate at a similar rate at various serum concentrations (Figure 3c). These results suggest that gag–RXR does not enhance the rate of cell proliferation, but rather blocks TGF-- or RA-induced growth inhibition.

PLZF–RAR fusion protein inhibits TGF--induced growth inhibition

While gag–RXR is an artificial fusion protein, the partner of RXR, RAR, is well known to be fused to other genes in human leukemic cells because of chromosomal translocations. Two of these oncogenic fusion proteins, PML–RAR and PLZF–RAR, have been reported to play an important role in the development of human APL (Look, 1997; Slack, 1999; Pandolfi, 2001). Although RAR and RXR are different genes, they do interact with each other and form a heterodimer that mediates RA-triggered gene transcription (Mangelsdorf and Evans, 1995). It is thus possible that PML–RAR and PLZF–RAR mimic the role of gag–RXR and also cause resistance to TGF--induced growth inhibition in mink lung epithelial cells. To test this hypothesis, we generated retroviral constructs that express each of the following proteins: RAR, truncated RAR (T-RAR), PML–RAR, and PLZF–RAR (Figure 4a). The expression of these proteins was confirmed by transient transfection of E-NX cells and subsequent immunoblotting analysis for RAR (Figure 4b, top panel, lane 3), PML–RAR (middle panel, lane 4), and PLZF–RAR (bottom panel, lane 5). The size for truncated RAR (45 kD) overlaps with a crossreactive band and thus is not distinctively identified (top panel, lane 2).

Figure 4.

PLZF–RAR and PML–RAR fusion proteins antagonize TGF--induced growth inhibition. (a) A diagram of constructs for RAR and its fusion proteins. (b) Expression of RAR and its fusion proteins in transfected E-NX cells. On day 0, 8x10⁵ E-NX cells were seeded in each well of six-well plates. On day 1, the cells were transfected with 2.5 g of the indicated DNAs. Nuclear extracts (90 g) were resolved on SDS–PAGE, and then processed for immunoblotting with each of the following antibodies: polyclonal anti-RAR antibody (Oncogene Research Products), anti-PML antibody (Santa Cruz Biotech, PG-m3), and anti-c-myc epitope antibody (Babco, 9E10), as described in Materials and methods. The asterisks (*) denote nonspecific crossreactive bands. However, the predicted size for T-RAR overlaps with the crossreactive band. (c) PLZF–RAR and PML–RAR cause resistance to TGF--mediated growth inhibition. ML-ER5 cells infected with control viruses (vector) or viruses expressing RAR, truncated RAR, PML–RAR or PLZF–RAR as indicated were seeded in 100 mm dishes at a density of 510⁵ cells/dish on day 0. On day 1, TGF- was added to each dish at a concentration of 30 pM. The medium was changed every 4 days. Cells were stained on day 13 with crystal violet for living cells

Full figure and legend (105K)

The recombinant retroviruses for each of the above constructs (Figure 4a) were packaged and introduced into ML-ER5 cells by infection. Infected cells were then subjected to selection by TGF-, followed by crystal violet staining for living cells. Figure 4c shows that cells infected with control viruses, or viruses carrying either RAR or truncated RAR, failed to grow in the presence of TGF-. However, cells infected with either PML–RAR or PLZF–RAR become resistant to TGF--induced growth inhibition, and continue to proliferate in the presence of TGF- (Figure 4c). Based on the staining, PLZF–RAR appeared to be more potent than PML–RAR in rendering cells resistant to TGF- (Figure 4c), and thus we chose to further analyze proliferation of cells expressing PLZF–RAR.

Fusion of PLZF and RAR is essential for resistance to TGF--induced growth inhibition

To test if the fusion of PLZF and RAR, like gag–RXR, is essential for resistance to TGF-, we generated various retroviral constructs that express the following proteins: PLZF–RAR, T-PLZF–RAR (with the POZ and proline-rich regions in PLZF truncated), PLZF and T-RAR (Figure 5a). Expression of these proteins with the predicted size was confirmed by Western blot analysis of transfected cells (Figure 5b). The resulting recombinant retroviruses, as indicated in Figure 5a, were used to infect ML-ER5 cells, followed by TGF- selection and crystal violet staining for living cells. Figure 5c shows that only PLZF–RAR, but not T-PLZF–RAR, PLZF or T-RAR, renders cells resistant to TGF-. These results suggest that the fusion between PLZF and RAR is essential for resistance to TGF-, and the POZ domain and the proline-rich region in PLZF play an important role in causing cells to be resistant to TGF-.

Figure 5.

Fusion of PLZF and RAR is essential for resistance to TGF-. (a) A diagram of various constructs involving PLZF and RAR. All the coding regions were tagged with c-Myc epitope at the NH-terminus. (b) Immunoblot analysis of expression of the proteins described in (a) with anti-c-Myc antibody. (c) Crystal violet staining of living cells that are resistant to TGF--induced growth inhibition. Cells were set up, treated, and stained as described in Figure 2d

Full figure and legend (94K)

RA potentiates the TGF--induced inhibition of cell proliferation and TGF--induced transcription of p15^ink4b

The observation that gag–RXR confers resistance to growth inhibition induced by either TGF- or RA (Figure 3) suggests a possible cooperation between TGF- and RA in lung epithelial cells. To explore this possibility, we treated mink lung epithelial cells with either TGF- or RA alone, or with a combination of TGF- and RA, and then determined the incorporation of ³H-thymidine into the treated cells. Figure 6a shows that TGF- and RA cooperatively inhibit proliferation of the lung epithelial cells.

Figure 6.

Cooperation of TGF- and RA in inhibition of cell growth and in induction of p15^ink4b. (a) On day 0, ML-ER5 cells infected with control viruses were seeded in 96-well plates at 510³ cells/200 l medium/well. TGF- (2 l in normal medium), RA (2 l diluted in normal medium), or both were added to each well to achieve final concentrations of 2 pM and 10^-8 M, respectively. RA was initially dissolved in ethanol. Determination of ³H-thymidine incorporation was performed as described in Experimental procedures. The difference in ³H-thymidine incorporation is highly significant according to Student's t-test between the TGF-+RA-treated cells and the control cells (P=0.002), between the TGF- +RA-treated cells and TGF--treated cells (P=0.0004), and between the TGF-+RA-treated cells and RA-treated cells (P=0.003). This is a representative of three independent experiments. (b) Cooperation in induction of p15^ink4b by TGF- and RA in lung epithelial cells. On day 0, ML-ER5 cells were seeded in 150 mm dishes (710⁶ cells per dish). On day 1, TGF- (200 pM), RA (10 M), or both were added to each dish in a staggering manner, and further incubated for the indicated periods of time. After treatment, the cells were harvested for RNA preparation. RNA (33 g per lane) was loaded into each lane, and the blot was probed with a fragment of cDNA for p15^ink4b. The middle panel is the membrane hybridized to the GAPDH probe. In the bottom panel, the signals for p15^ink4b and GAPDH were quantitated using the NIH Image software, and the values are the ratio of the p15^ink4b signal and the GAPDH signal. This is a representative of two independent experiments

Full figure and legend (116K)

RA induces growth inhibition in a variety of cells by inducing production of active TGF-, leading to growth inhibition (Roberts and Sporn, 1992; Nunes et al., 1996; Defacque et al., 1999; Borger et al., 2000). p15^ink4b is an inhibitor of cyclin-dependent kinase 4/6 (CDK4/6) and is potently induced by TGF- (Hannon and Beach, 1994), but it is not clear whether this gene is also induced by RA. We tested whether TGF- and RA cooperatively induce transcription of various growth-inhibiting genes including p15^ink4b (Figure 6b) and found that TGF- alone induces transcription of p15^ink4b, which peaks 5 h after stimulation (Figure 4b, lane 4), consistent with previous reports (Hannon and Beach, 1994). RA also stimulates transcription of p15^ink4b, which peaks 2 h after stimulation (lane 7), albeit to a lesser extent compared to the induction by TGF-. Stimulation of the mink lung cells with both TGF- and RA cooperatively induces transcription of p15^ink4b, which peaks at 5 h after the costimulation (lane 12). Since RA only modestly induces p15^ink4b, it may mainly potentiate TGF--mediated induction of the p15^ink4b gene. Together, these results suggest that TGF- and RA induce growth inhibition of mink lung epithelial cells in part through induction of p15^ink4b transcription.

Expression of PLZF–RAR represses induction of transcription by TGF-

Since PLZF–RAR antagonizes TGF--induced growth inhibition, we examined whether PLZF–RAR affects the TGF--induced translocation of Smad2/Smad3 into the nucleus. ML-ER5 control cells or the cells expressing PLZF–RAR were treated with or without TGF-, and the resulting nuclear extracts were immunoblotted using an anti-Smad2/Smad3 antibody. The results show that Smad2/Smad3 in both the control cells and the PLZF–RAR-expressing cells translocate into the nucleus after TGF- treatment (Figure 7a). A synthetic promoter with two RA response elements (RAREs) and two Smad binding elements (SBEs) was used to drive the expression of a luciferase reporter gene (pRS-Luc) (Figure 7b). HepG2 cells were cotransfected with pRS-Luc and a control vector or PLZF–RAR, and the transfected cells were induced with TGF- and/or RA, followed by luciferase assays. Figure 7b shows that TGF- and RA cooperatively induce expression of the pRS-Luc reporter gene while PLZF–RAR suppresses the TGF--induced transcription. Together, these results suggest that PLZF–RAR does not block TGF--induced translocation of Smad2/Smad3, but rather antagonizes a step of TGF--regulated transcription.

Figure 7.

PLZF–RAR-mediated resistance to TGF--induced transcription. (a) PLZF–RAR does not block TGF--induced translocation of Smad2 and Smad3 proteins. On day 0, 410⁶ control cells and ML-ER5 cells expressing PLZF–RAR were seeded in a 150 mm dish. On day 1, cells were treated with 200 pM TGF- for 1 h before harvesting for preparation of nuclear extracts. For immunoblotting analysis, 35 g of protein from nuclear extract was loaded in each lane for separation, and blotted with the antibody against Smad3 and Smad2 (Santa Cruz Biotechnology #sc-8332). (b) PLZF–RAR inhibits TGF--induced transcription. On day 0, 20⁵ HepG2 cells were seeded per well in 12-well plates. On day 1, cells in each well were transfected with 0.125 g of pCMV--gal and 2.5 g of pRS-Luc, together with 0.45 g of pEXL-GFP (control) or pEXL-PLZF–RAR (PLZF–RAR) as indicated, by the calcium phosphate precipitation method. On day 2, cells were switched to normal medium. On day 3, transfected cells were treated with or without 200 pM TGF- and/or 10^-5 M RA in DMEM, as indicated. After 24 h of incubation, cells were harvested for luciferase and -galactosidase assays. The difference in luciferase activities is significant according to Student's t-test between the TGF--treated cells and the TGF-+RA-treated cells in the absence of PLZF–RAR (P=0.008), between the TGF--treated cells in the absence or presence of PLZF–RAR (P=0.01), and between the TGF-+RA-treated cells in the absence or presence of PLZF–RAR (P=0.045), as indicated. This is a representative of three independent experiments

Full figure and legend (93K)

Expression of PLZF–RAR attenuates TGF--induced expression of p15^INK4b

Since TGF- induces transcription of p15^ink4b (Hannon and Beach, 1994), we examined the effect of expression of PLZF–RAR on the TGF--mediated regulation of the p15^INK4b gene by Northern blot analysis. The results show that TGF- induced accumulation of the p15^INK4b mRNA in control cells (Figure 8, top panel, lanes 1–4). In contrast, induction of the p15^INK4b mRNA diminishes by 50% in the cells expressing PLZF–RAR at a concentration of 40 pM of TGF- (Figure 8, lanes 3 and 7). At a concentration of 200 pM TGF-, PLZF–RAR reduces the level of the p15^ink4b mRNA by 40% (lanes 4 and 8). The signal of GAPDH (for loading control) (middle panel) and the normalized p15^INK4b signal (bottom panel) are also shown in Figure 8. Together, these results suggest that PLZF–RAR represses TGF--induced transcription of the p15^ink4b gene.

Figure 8.

PLZF–RAR antagonizes TGF--mediated regulation of p15^INK4b. Northern blot analysis of p15^INK4b in control cells and PLZF–RAR-expressing cells. Top panel, control ML-ER5 cells containing a vector (lanes 1–4) or PLZF–RAR (lanes 5–8) were cultured in the presence of 0, 8, 40, or 200 pM TGF- for 14 h. Total RNA was isolated from the treated cells and hybridized to a p15^INK4b probe corresponding to the first exon of the human p15^INK4b gene as described in Materials and methods. Middle panel shows the membrane filter hybridized with the control GAPDH probe. Bottom panel, the bar graph shows the relative quantity of the p15^INK4b mRNA in each lane. The signal for each lane in the membrane filter was detected on a Molecular Dynamics Phosphorimager, quantitated with Probe Quant software, and normalized with the corresponding GAPDH signal. This is a representative of two independent experiments

Full figure and legend (98K)

Smad proteins interact with RAR and PLZF–RAR

We examined the possible association between Smad3 and RAR or PLZF–RAR, and whether the interaction is regulated by TGF--induced phosphorylation of Smad3. Human 293 cells were transfected with cDNAs for RAR or PLZF–RAR together with the Flag epitope-tagged Smad3. Constitutively active TGF- type I receptor, TRI-T204D, was also included in some transfections because it is well known to actively phosphorylate Smad2 and Smad3 proteins, mimicking the function of TGF-. Cell lysates from transfected cells were incubated with agarose beads conjugated with anti-Flag epitope antibodies. Flag-tagged Smad3 and its associated proteins were immunoprecipitated with the conjugated agarose beads, separated on a polyacrylamide gel, and immunoblotted with an anti-RAR antibody. Figure 9 (top panel) shows that Smad3 interacts with RAR only in the presence of constitutively active TRI (lane 3) but not in the absence of the active receptor (lane 2), suggesting that the Smad3 and RAR interaction is specific. Similarly, PLZF–RAR interacts with Smad3 only in the presence of the active TRI (lane 6), but not in the absence of the active receptor (lane 5). The input of RAR/PLZF–RAR and Smad3 is similar (middle and bottom panel). Together, these results indicate that TGF--induced phosphorylation of Smad3 enhances its interaction with RAR and PLZF–RAR inside cells, suggesting that the interaction between Smad3 and RAR or PLZF–RAR is specific. However, it cannot be ruled out that the constitutively active type I receptor also induces secretion of other factors that may ultimately modulate TGF--regulated interaction between Smad3 and RAR or PLZF–RAR, as a recent report has shown that TGF- induces secretion of fibulin 5, which regulates several TGF--regulated biological activities (Schiemann et al., 2002).

Figure 9.

Interaction between Smad3 and RAR or PLZF–RAR in cells. On day 0, 293 cells were seeded at a density of 2 10⁶ cells/60 mm dish. On day 1, cells were transfected, as described in Materials and methods,with 2.5 g of Flag epitope-tagged Smad3 (Hua et al., 1998) and 2.5 g of RAR (pEXL-RAR) or PLZF–RAR (pEXL-PLZF–RAR), with or without a constitutively active TGF- type I receptor, T204D (1.0 g), as indicated. The total amount of DNA for transfection per dish was adjusted to 6.0 g with pcDNA3. Transfected cells were switched to fresh medium on day 2 and harvested on day 3. Cell lysates were prepared, and used for immunoprecipitation as described in Materials and methods. The Smad3-associated proteins on the M2 beads conjugated with an anti-Flag antibody (Sigma) were separated on 8% SDS–PAGE gels, and blotted with an anti-RAR polyclonal antibody (Oncogene Research Products, PC92L) as described in Materials and methods. The exposure time for the top panel was 20 s. To confirm that the expression of RAR, PLZF–RAR, and Smad3 is as expected, 25 g of proteins from each group of transfected cells were separated by SDS–PAGE, and detected by an anti-RAR antibody (middle panel), or by an anti-Flag antibody (bottom panel). This is a representative of two independent experiments

Full figure and legend (82K)

Discussion

In screening for cDNAs whose expression causes resistance to TGF--induced growth inhibition, gag–RXR was identified as a fusion protein that antagonizes TGF--induced inhibition of cell proliferation (Figure 3). A naturally occurring oncogene, v-ERBA, is also a product of the fusion between the retroviral gag protein and the chicken thyroid hormone receptor. v-ERBA inhibits the transcriptional activation by normal thyroid hormone receptors and transforms hematopoietic and nonhematopoietic cells (Thormeyer and Baniahmad, 1999). Thus, fusion with gag sequence may be an effective way to subvert the function of some steroid hormone receptors. Although gag–RXR is an artificial fusion protein, a partner of RXR, RAR, is a portion of several naturally occurring oncogenic fusion proteins such as PML–RAR and PLZF–RAR. These fusion proteins are formed because of chromosomal translocations and induce APL (Look, 1997; Slack, 1999; Pandolfi, 2001). The observations on gag–RXR prompted us to examine whether oncogenic PML–RAR and PLZF–RAR also cause resistance to TGF--induced inhibition of cell growth. Notably, the results demonstrate that both PML–RAR and PLZF–RAR indeed antagonize TGF--induced growth inhibition in mink lung epithelial cells (Figure 4).

Although previous reports have shown that PLZF–RAR inhibits RA signaling by blocking the RAR-mediated transcriptional activation (Slack, 1999; Pandolfi, 2001), the current studies demonstrate that PLZF–RAR compromises TGF--induced growth inhibition. PLZF–RAR does not block TGF--induced translocation of Smad2 and Smad3 into the nucleus (Figure 7a), but rather inhibits TGF--induced expression of a reporter gene (Figure 7b) as well as the endogenous p15^ink4b gene (Figure 8). These results suggest that PLZF–RAR antagonizes TGF- signaling at a step of the TGF--induced gene transcription downstream of Smad signaling, but not the TGF--induced nuclear translocation of Smad proteins.

Fusion between PLZF and RAR is essential for PLZF–RAR in antagonizing TGF- signaling (Figure 5). It has previously been reported that PLZF–RAR binds RAREs and recruits transcriptional repressors including N-Cor/SMRT, mSin3, and histone deacetylases leading to inhibition of gene transcription (Slack, 1999). Since PLZF and RAR each contain a domain binding these repressors, a fusion between these two proteins may generate a protein with more repressor-binding domains and thus enhanced strength in repressing gene transcription (Slack, 1999). Coupled with the result that PLZF–RAR selectively interacts with activated Smad3 (Figure 9), it is possible that PLZF–RAR may actively recruit transcriptional corepressors to repress growth-inhibiting genes that are normally induced by Smad proteins and RAR. Hence, the current findings suggest that PLZF–RAR may induce APL (Ruthardt et al., 1997; He et al., 2000) because of resistance not only to RA but also to TGF--induced growth inhibition, providing a novel mechanism of tumorigenesis mediated by PLZF–RAR.

It has previously been reported that TGF- and RA can cooperate to inhibit cell proliferation (Roberts and Sporn, 1992; Dickens and Colletta, 1993). The prevalent model suggests that RA inhibits a variety of cells partly by stimulating the production of active TGF- (Nunes et al., 1996; Defacque et al., 1999; Borger et al., 2000), since a neutralizing anti-TGF- antibody blocks RA's ability to inhibit cell growth. However, since TGF- was inactivated by the neutralizing anti-TGF- antibody in the culture medium in these reports, the possible intracellular cooperation between RA and TGF- could not be addressed. Our findings that gag–RXR causes resistance to TGF- as well as RA raises the possibility of the intracellular cooperation between the downstream signaling components of the TGF- and RA pathways. Supporting this hypothesis, Smad3 interacts with RAR and PLZF–RAR in a TGF--dependent manner (Figure 9), and PLZF–RAR inhibits TGF--induced transcription of p15^ink4b, likely through transcriptional regulation (Figure 8). Together, these data suggest the existence of intracellular cooperation between the TGF- and RA signaling pathways, extending the current model about the interaction between these two pathways.

The current studies have shown that TGF- and RA cooperatively inhibit proliferation of mink lung epithelial cells (Figure 6a) and RA markedly potentiates TGF--induced transcription of p15^ink4b (Figure 6b). However, it is also likely that other genes in addition to p15^ink4b are also involved in the cooperation between the TGF- and RA pathways. For example, RA induces ubiquitin-dependent degradation of cyclin D1, a critical component of CDK4/6 that can be repressed by TGF- signaling, in a carcinoma cell line (Spinella et al., 1999).

Although TGF- and RA cooperatively inhibit cell proliferation and gene transcription in the lung epithelial cells, RAR may not be required for TGF- signaling, but instead, it may potentiate the TGF- signaling pathway in the presence of RA. It is likely that TGF- and RA each signal through its own signaling pathway to inhibit cell proliferation. The promoter of some genes may contain both TGF- and RA response elements and transcription of these genes may be cooperatively induced by TGF- and RA.

The current studies, beginning from an unexpected finding that the gag–RXR fusion protein interferes with TGF--induced growth inhibition, lead to unraveling the role of PLZF–RAR in resisting TGF--induced growth inhibition. These findings not only suggest the existence of intracellular cooperation between the TGF- and RA signaling pathways, but also provide new insights into a likely link between resistance to TGF- and the development of the PLZF–RAR-induced leukemia.

Materials and methods

Plasmid construction

Standard molecular biology techniques were used as described (Sambrook et al., 1989). Oligonucleotides were synthesized by Integrated DNA Technologies, Inc. To clone RXR cDNA into the retroviral vector pMX-Puro (Onishi et al., 1996), mink RXR from TGF--resistant clone 19, which lacks the first seven amino-acid residues, was amplified by PCR using pfu polymerase. The amplified RXR was digested and cloned into the BamHI and NotI site of pMX-Puro to generate pMX-RXR. To prepare constructs expressing myc-tagged RXR- or gag–RXR, a pair of oligonucleotides encoding the myc-epitope tag was synthesized and inserted into the BamHI and XhoI site of pMX-puro to generate pMyc-MP. Pairs of oligonucleotides were used to amplify RXR and gag–RXR by PCR from clone 19, and the PCR products were digested with MluI/XhoI and MluI/NotI, respectively, and then inserted into the corresponding site in pMyc-MP to generate pMyc-RXR and pMyc-gag–RXR, respectively. To construct various insertional mutants of gag–RXR as indicated in Figure 2b, various oligonucleotides that carry different inserts between the gag and the RXR- sequences were used to amplify RXR from clone 19. PCR products were digested with BglII and NotI, and the resulting digested fragments were cloned into the BamHI and NotI site of pMX-Puro to generate the various insertional mutants as shown in Figure 2b. To express RXR without fusion to gag, an oligonucleotide carrying a stop codon followed by the ATG codon and the eighth amino-acid residue of mink RXR was used to amplify RXR from clone 19. The PCR product was then digested with BglII and NotI, and inserted into the BamHI and NotI site of pMX-Puro to generate pM-RXR.

To express RAR, T-RAR, and PML–RAR in a retroviral vector, pairs of oligonucleotides that correspond to the 5' and 3' ends of each of the above cDNAs were used to PCR-amplify the above cDNAs from pCMX-RAR, pSG5-RAR, and pSG5-PML–RAR, respectively. The resulting PCR products were digested with BamHI and NotI, and then inserted into the BamHI and NotI site in pMX-Puro to generate the corresponding constructs as indicated in Figure 4a. Additionally, PLZF–RAR was PCR amplified from pcDNA3-PLZF–RAR and then cloned into the MluI–NotI site of pmyc-MP to generate pmyc-PLZF–RAR. To construct pEXL-RAR and pEXL-PLZF–RAR, RAR and PLZF–RAR were released with BamHI and NotI from pRAR and pMyc-PLZF–RAR, and then inserted into the BamHI and NotI sites of pEXL-GFP (Liu et al., 1997a). To construct pRS-Luc, a pair of oligonucleotides containing two direct repeats of a RARE (AAGGGTTCACCGAAAGTTCACTCGCAT) (Umesono et al., 1991) was inserted into the EcoRI/KpnI site of a modified pE2.1-Luc (Hua et al., 1999). Thus, pRS-Luc contains a synthetic promoter that consists of two direct repeats of a RARE and two repeats of a Smad binding element. All of the above constructs were sequenced to verify the junction and mutation, and two independent clones were routinely tested in each experiment. Additionally, PLZF–RAR, T-PLZF–RAR, PLZF, and T-RAR were PCR-amplified from pcDNA3-PLZF–RAR and then cloned into the MluI-NotI site of pMyc-MP to generate the constructs with a c-myc epitope as indicated in Figure 5a.

Tissue culture

Mink lung epithelial cells (MvlLu from ATCC) and their derivative cells were cultured in MEM supplemented with 10% fet al bovine serum, 1% MEM nonessential amino-acid mixture, 100 U/ml penicillin and 100 g/ml streptomycin. To generate a mink lung epithelial cell line that could be infected by murine Moloney leukemia retroviruses, the cDNA for the ecotropic receptor for murine Moloney leukemia retrovirus was transfected into MvlLu cells. The resulting cells, ML-ER5, were able to be infected efficiently by recombinant retroviruses (Hua et al., 1998). Human embryonic kidney (HEK) 293 cells, human hepatoma HepG2 cells, and retroviral packaging cells (Bosc23 cells and E-NX cells) were cultured in DMEM containing 10% fet al bovine serum, 100 U/ml penicillin, and 100 g/ml streptomycin. All of the above cells were cultured in 5% CO₂ at 37°C. TGF-1 was from R&D systems, and all trans- RA was from Sigma and Aldrich, Inc. Treatment of cells with TGF- and/or RA was accomplished by direct addition to normal medium except for luciferase assays, for which DMEM without any supplements was used.

Construction of retroviral cDNA library, infection of cells, and selection for TGF--resistant cells

A retroviral cDNA library was made as previously described (Hua et al., 1998). Briefly, poly (A)⁺ RNA was isolated from MvlLu cells. cDNAs were synthesized from the poly (A)⁺ RNA using the random-primer method, and then cloned into the EcoRI site of the retroviral vector pMX. The resulting cDNA library was introduced into the Bosc23 packaging cell line (Pear et al., 1993) to obtain a high titer retroviral cDNA library.

To select for TGF--resistant cell clones, ML-ER5 cells (10⁶ cells/150 mm dish) were infected with control or cDNA-carrying recombinant retroviruses (1 : 3 dilution) for 9 h, and then switched to normal medium. After 24 h, infected cells were subjected to selection with 10 pM of TGF-. On day 12 after infection, TGF--resistant clones were isolated and then expanded. The cDNAs responsible for resistance to TGF- were isolated as previously described (Hua et al., 1998).

Transfection and luciferase assays

A total of 293 cells, HepG2 cells, and E-NX cells were transfected by the calcium phosphate precipitation method (Sambrook et al., 1989). For luciferase assays, cells were cotransfected with pCMV- encoding the lacZ gene (Clontech) as an internal control to normalize the luciferase activity. To transfect HepG2 cells, cells were seeded at a density of 210⁵ cells per well in 12-well plates unless otherwise stated. On day 1, cells were switched to fresh medium and then transfected by the calcium phosphate precipitation method. After overnight incubation, transfected cells were switched to normal medium. On day 3, cells were switched to serum-free medium, and TGF- and/or RA was added to the medium. After 24 h of induction, cells were harvested for luciferase and -galactosidase assays. Both the luciferase and -galactosidase activities were measured by a TR717 Microplate Luminometer (EG&G Berthold). All luciferase activities were normalized to the -galactosidase activities and presented as an average of duplicate samples as previously described (Hua et al., 1998).

³H-thymidine incorporation assay

On day 0, 5000 cells in 200 l of normal medium were seeded in each well of a 96-well plate. Cells were treated using various concentrations of TGF- and/or RA and incubated in the presence of TGF- for 4 days. On day 4, ³H-thymidine (ICN Biomedicals Inc) was added to the medium. ³H-thymidine was first diluted in Opti-MEM (Gibco-BRL) and then was added to each well. Duplicate samples were used. Cells were incubated at 37°C, 5% CO₂ for 4 h before being trypsinized and harvested to a 96-well filtermat for Model MK111 Harvester 96 (Tomtec). Incorporation of ³H-thymidine into cells was determined using a Model 1450 Microbeta Trillux Scintillation Counter (Wallac).

Northern blot analysis

ML-ER5 cells were seeded in 150 mm dishes at a density of 710⁶ cells in 20 ml of normal medium. After 48 h, fresh medium was added together with 200 pM TGF- and/or 10 M RA at staggered time intervals. Cells were harvested and total RNA was isolated using the cesium chloride centrifugation method (Hua et al., 1996). RNA (33 g) was resolved on a 1% agarose gel with formaldehyde (Sambrook et al., 1989), and then transferred overnight by capillary action to Hybond-N⁺ membrane (Amersham-Pharmacia). A cDNA probe from the 5' untranslated region of p15^ink4b was generated by EcoRI digestion of pCRII-p15. The p15 probe was labeled with ³²P--dCTP using a Ready-To-Go-DNA Labeling kit (Amersham-Pharmacia). The probe was hybridized to the membrane in Quik-hyb solution (Stratagene) as instructed by the manufacturer.

Cell fractionation, immunoprecipitation, and immunoblotting

Nuclear extracts from transfected cells were prepared as previously described (Hua et al., 1995). To prepare whole cell lysate, transfected 293 cells or E-NX cells were harvested and lysed in NP-40 lysis buffer (30 mM Tris, pH 8.0, 100 mM NaCl, 25 mM -glycerophosphate, 1% NP-40, and fresh protease cocktails from CalbioChem., Inc.). The cells were rotated at 60 r.p.m. at 4°C for 2 h. The supernatant (350 l, 1.25 mg of proteins) from the cell lysates was supplemented with 12% glycerol, and then mixed with 20 l of equilibrated anti-Flag antibody-conjugated agarose beads (M2 beads, 50% slurry, Sigma Inc.). The cell lysates and M2 beads were rotated at 30 r.p.m. at 4°C for 4 h. Then the beads were centrifuged and washed 5 times in washing buffer (30 mM Tris, pH 8.0, 150 mM NaCl, 25 mM glycerophosphate, 1% NP-40, and 1x fresh protease inhibitor cocktails from CalbioChem. Inc.). The washed beads and cell lysates were separated on SDS–PAGE, and blotted with the indicated antibodies, and the proteins were detected using enhanced Chemiluminescence reagents from Amersham-Pharmacia, Inc.

References

1. Attisano L and Wrana JL. (2000). Curr. Opin. Cell. Biol., 12, 235–243. | Article | PubMed | ISI | ChemPort |
2. Batova A, Danielpour D, Pirisi L and Creek KE. (1992). Cell Growth Differ., 3, 763–772.
3. Blobe GC, Liu X, Fang SJ, How T and Lodish HF. (2001). J. Biol. Chem., 276, 39608–39617. | Article | PubMed | ISI | ChemPort |
4. Borger DR, Mi Y, Geslani G, Zyzak LL, Batova A, Engin TS, Pirisi L and Creek KE. (2000). Virology, 270, 397–407. | PubMed |
5. Bottinger EP, Jakubczak JL, Roberts ISD, Mumy M, Hemmati P, Bagnall K, Merlino G and Wakefield LM. (1997). EMBO J., 16, 2621–2633. | Article | PubMed | ISI | ChemPort |
6. Chen X, Rubock MJ and Whitman M. (1996). Nature, 383, 691–696. | Article | PubMed | ISI | ChemPort |
7. Chipuk JE, Cornelius SC, Pultz NJ, Jorgensen JS, Bonham MJ, Kim SJ and Danielpour D. (2002). J. Biol. Chem., 277, 1240–1248. | Article | PubMed | ISI | ChemPort |
8. Danielpour D. (1996). J. Cell Physiol., 166, 231–239. | Article | PubMed | ISI | ChemPort |
9. Datta PK, Blake MC and Moses HL. (2000). J. Biol. Chem., 275, 40014–40019. | Article | PubMed | ISI | ChemPort |
10. Datto MB, Frederick JP, Pan L, Borton AJ, Zhuang Y and Wang XF. (1999). Mol. Cell Biol., 19, 2495–2504. | PubMed | ISI | ChemPort |
11. Defacque H, Piquemal D, Basset A, Marti J and Commes T. (1999). J. Cell Physiol, 178, 109–119. | Article | PubMed | ChemPort |
12. Derynck R, Zhang Y and Feng XH. (1998). Cell, 95, 737–740. | Article | PubMed | ISI | ChemPort |
13. Dickens TA and Colletta AA. (1993). Bioessays, 15, 71–74.
14. Dragnev KH, Freemantle SJ, Spinella MJ and Dmitrovsky E. (2001). Ann. NY Acad. Sci., 952, 13–22. | PubMed | ChemPort |
15. Eppert K, Scherer SW, Ozcelik H, Pirone R, Hoodless P, Kim H, Tsui LC, Bapat B, Gallinger S, Andrulis IL, Thomsen GH, Wrana JL and Attisano L. (1996). Cell, 86, 543–552. | Article | PubMed | ISI | ChemPort |
16. Falk LA, De Benedetti F, Lohrey N, Birchenall-Roberts MC, Ellingsworth LW, Faltynek CR and Ruscetti FW. (1991). Blood, 77, 1248–1255. | PubMed | ISI | ChemPort |
17. Feng X, Zhang Y, Wu R and Derynck R. (1998). Genes Dev., 15, 2153–2163.
18. Feng XH, Liang YY, Liang M, Zhai W and Lin X. (2002). Mol. Cell, 9, 133–143. | Article | PubMed | ISI | ChemPort |
19. Feng XH, Lin X and Derynck R. (2000). EMBO J., 19, 5178–5193. | Article | PubMed | ISI | ChemPort |
20. Geradts J, Chen JY, Russell EK, Yankaskas JR, Nieves L and Minna JD. (1993). Cell Growth Differ., 4, 799–809. | PubMed |
21. Glick AB, Flanders KC, Danielpour D, Yuspa SH and Sporn MB. (1989). Cell Regul., 1, 87–97. | PubMed | ISI | ChemPort |
22. Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I, Seiser C, Lazar MA, Minucci S and Pelicci PG. (1998). Nature, 391, 815–818. | Article | PubMed | ISI | ChemPort |
23. Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A. (1998). Blood, 91, 2634–2642. | PubMed | ISI | ChemPort |
24. Hahn SA, Schutte M, Hoque ATMS, Moslaluk CA, da Costa LT, Rozenblum E, Weinstein CL, Fischer A, Yeo cJ, Hruban RH and Kern SE. (1996). Science, 271, 350–353. | Article | PubMed | ISI | ChemPort |
25. Han GR, Dohi DF, Lee HY, Rajah R, Walsh GL, Hong WK, Cohen P and Kurie JM. (1997). J. Biol. Chem., 272, 13711–13716. | Article | PubMed | ISI | ChemPort |
26. Hannon G and Beach D. (1994). Nature, 371, 257–261. | Article | PubMed | ISI | ChemPort |
27. Hayes SA, Zarnegar M, Sharma M, Yang F, Peehl DM, ten Dijke P and Sun Z. (2001). Cancer Res., 61, 2112–2118. | PubMed | ISI | ChemPort |
28. He L, Bhaumik M, Tribioli C, Rego EM, Ivins S, Zelent A and Pandolfi PP. (2000). Mol. Cell, 6, 1131–1141. | Article | PubMed | ISI | ChemPort |
29. He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A and Pandolfi PP. (1998). Nat. Genet., 18, 126–135. | Article | PubMed | ISI | ChemPort |
30. Hong SH, David G, Wong CW, Dejean A and Privalsky ML. (1997). Proc. Natl. Acad. Sci. USA, 94, 9028–9033. | Article | PubMed | ChemPort |
31. Hua X, Liu X, Ansari DO and Lodish HF. (1998). Genes Dev., 12, 3084–3095. | PubMed | ISI | ChemPort |
32. Hua X, Miller ZA, Wu G, Shi Y and Lodish HF. (1999). Proc. Natl. Acad. Sci. USA, 96, 13130–13135. | Article | PubMed | ChemPort |
33. Hua X, Nohturft A, Goldstein JL and Brown MS. (1996). Cell, 87, 415–426. | Article | PubMed | ISI | ChemPort |
34. Hua X, Sakai J, Ho YK, Goldstein JL and Brown MS. (1995). J Biol. Chem., 270, 29422–29427. | Article | PubMed | ISI | ChemPort |
35. Imai S, Okuno M, Moriwaki H, Muto Y, Murakami K, Shudo K, Suzuki Y and Kojima S. (1997). FEBS Lett., 411, 102–106. | Article | PubMed | ISI | ChemPort |
36. Imai Y, Kurokawa M, Izutsu K, Hangaishi A, Maki K, Ogawa S, Chiba S, Mitani K and Hirai H. (2001). Oncogene, 20, 88–96. | Article | PubMed | ISI | ChemPort |
37. Jakubowiak A, Pouponnot C, Berguido F, Frank R, Mao S, Massague J and Nimer SD. (2000). J. Biol. Chem., 275, 40282–40287. | Article | PubMed | ISI | ChemPort |
38. Janknecht R, Wells NJ and Hunter T. (1998). Genes Dev., 12, 2114–2119. | PubMed | ISI | ChemPort |
39. Kim RH, Wang D, Tsang M, Martin J, Huff C, de Caestecker MP, Parks WT, Meng X, Lechleider RJ, Wang T and Roberts AB. (2000). Genes Dev., 14, 1605–1616. | PubMed | ISI | ChemPort |
40. Kishi H, Kuroda E, Mishima HK and Yamashita U. (2001). Cell Biol. Nit., 25, 1125–1129.
41. Kojima S and Rifkin DB. (1993). J. Cell Physiol., 155, 323–332. | PubMed |
42. Kurokawa M, Mitani K, Irie K, Matsuyama T, Takahashi T, Chiba S, Yazaki Y, Matsumoto K and Hirai H. (1998). Nature, 394, 92–96. | Article | PubMed | ISI | ChemPort |
43. Lagna G, Hata A, Kemmati-Brivanlou A and Massague J. (1996). Nature, 383, 832–836. | Article | PubMed | ISI | ChemPort |
44. Lardon F, Snoeck HW, Haenen L, Lenjou M, Nijs G, Weekx SF, Van Ranst PC, Berneman ZN and Van Bockstaele DR. (1996). Leukemia, 10, 1937–1943. | PubMed |
45. Li J-M, Nichols MA, Chandrasekharan S, Xiong Y and Wang X-F. (1995). J. Biol. Chem., 270, 26750–26753. | Article | PubMed | ISI | ChemPort |
46. Liberati NT, Datto MB, Frederick JP, Shen X, Wong C, Rougier-Chapman EM and Wang XF. (1999). Proc. Natl. Acad. Sci. USA, 96, 4844–4849. | Article | PubMed | ChemPort |
47. Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH and Evans RM. (1998). Nature, 391, 811–814. | Article | PubMed | ISI | ChemPort |
48. Liu X, Sun Y, Constantinescu SN, Karam E, Weinberg RA and Lodish HF. (1997a). Proc. Natl. Acad. Sci. USA, 94, 10669–10674. | Article | PubMed | ChemPort |
49. Liu X, Sun Y, Constantinescu SN, Karam E, Weinberg RA and Lodish HF. (1997b). Proc. Natl. Acad. Sci. USA, 94, 10669–10674. | Article | PubMed | ChemPort |
50. Look AT. (1997). Science, 278, 1059–1064. | Article | PubMed | ISI | ChemPort |
51. Luo K, Stroschein SL, Wang W, Chen D, Martens E, Zhou S and Zhou Q. (1999). Genes Dev., 13, 2196–2206. | Article | PubMed | ISI | ChemPort |
52. Mangelsdorf DJ and Evans RM. (1995). Cell, 83, 841–850. | Article | PubMed | ISI | ChemPort |
53. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW, Vogelstein B, Brattain M and Willson JKV. (1995). Science, 268, 1336–1338. | Article | PubMed | ISI | ChemPort |
54. Massague J, Blain SW and Lo RS. (2000). Cell, 103, 295–309. | Article | PubMed | ISI | ChemPort |
55. Meyskens Jr FL and Manetta A. (1995). Am. J. Clin. Nutr., 62, 1417S–1419S.
56. Morales TI and Roberts AB. (1992). Arch. Biochem. Biophys., 293, 79–84.
57. Moustakas A and Kardassis D. (1998). Proc. Natl. Acad. Sci. USA, 95, 6733–6738. | Article | PubMed | ChemPort |
58. Nunes I, Kojima S and Rifkin DB. (1996). Cancer. Res., 56, 495–499. | PubMed | ISI | ChemPort |
59. Okuno M, Kojima S, Akita K, Matsushima-Nishiwaki R, Adachi S, Sano T, Takano Y, Takai K, Obora A, Yasuda I, Shiratori Y, Okano Y, Shimada J, Suzuki Y, Muto Y and Moriwaki Y. (2002). Front. Biosci., 7, D204–D218. | Article | PubMed | ISI | ChemPort |
60. Okuno M, Moriwaki H, Imai S, Muto Y, Kawada N, Suzuki Y and Kojima S. (1997). Hepatology, 26, 913–921. | PubMed | ISI | ChemPort |
61. Onishi M, Kinoshita S, Morikawa Y, Shibuya A, Phillips J, Lanier LL, Gorman DM, Nolan GP, Atsushi M and Kitamura T. (1996). Exp. Hematol., 24, 324–329. | PubMed | ISI | ChemPort |
62. Pandolfi PP. (2001). Hum. Mol. Genet., 10, 769–775. | Article | PubMed | ISI | ChemPort |
63. Pear WS, Nolan GP, Scott ML and Baltimore D. (1993). Proc. Natl. Acad. Sci. USA, 90, 8392–8396. | Article | PubMed | ChemPort |
64. Pietenpol JA, Holt JT, Stein RW and Moses HL. (1990). Proc. Natl. Acad. Sci. USA, 87, 3758–3762. | Article | PubMed | ChemPort |
65. Pouponnot C, Jayaraman L and Massague J. (1998). J. Biol. Chem., 273, 22865–22868. | Article | PubMed | ISI | ChemPort |
66. Razani B, Zhang XL, Bitzer M, von Gersdorff G, Bottinger EP and Lisanti MP. (2001). J. Biol. Chem., 276, 6727–6738. | Article | PubMed | ISI | ChemPort |
67. Roberts AB and Sporn MB. (1992). Cancer Surv., 14, 205–220. | PubMed | ISI | ChemPort |
68. Ruthardt M, Testa U, Nervi C, Ferrucci PF, Grignani F, Puccetti E, Peschle C and Pelicci PG. (1997). Mol. Cell Biol., 17, 4859–4869. | PubMed | ISI | ChemPort |
69. Sambrook J, Fritsch E and Maniatis T. (1989). Molecular Cloning. Cold Spring Harbor Laboratory Press, New York.
70. Schiemann WP, Blobe GC, Kalume DE, Pandey A and Lodish HF. (2002). J. Biol. Chem., 277, 27367–27377. | Article | PubMed | ISI | ChemPort |
71. Shen X, Hu PP, Liberati NT, Datto MB, Frederick JP and Wang X-F. (1998). Mol. Biol. Cell, 9, 3309–3319. | PubMed | ISI | ChemPort |
72. Shi Y, Wang Y-F, Jayaraman L, Yang H, Massague J and Pavletich NP. (1998). Cell, 94, 585–594. | Article | PubMed | ISI | ChemPort |
73. Slack JL. (1999). Curr. Opin. Oncol., 11, 9–13. | PubMed |
74. Spinella MJ, Freemantle SJ, Sekula D, Chang JH, Christie AJ and Dmitrovsky E. (1999). J. Biol. Chem., 274, 22013–22018. | Article | PubMed | ISI | ChemPort |
75. Sun Y, Liu X, Eaton EN, Lane WS, Lodish HF and Weinberg RA. (1999). Mol. Cell, 4, 499–509. | Article | PubMed | ISI | ChemPort |
76. Takaku K, Oshima M, Miyoshi H, Matsui M, Seldin MF and Taketo MM. (1998). Cell, 92, 645–656. | Article | PubMed | ISI | ChemPort |
77. Ten Dijke P, Goumans MJ, Itoh F and Itoh S. (2002). J. Cell. Physiol., 191, 1–16. | Article | PubMed | ISI | ChemPort |
78. Thormeyer D and Baniahmad A. (1999). Int. J. Mol. Med., 4, 351–358. | PubMed | ISI | ChemPort |
79. Tsukazaki T, Chiang TA, Davison AF, Attisano L and Wrana JL. (1998). Cell, 95, 779–791. | Article | PubMed | ISI | ChemPort |
80. Umesono K, Murakami K, Thompson C and Evans R. (1991). Cell, 5, 1255–1266.
81. Valette A and Botanch C. (1990). Growth Factors, 2, 283–287.
82. Virmani AK, Rathi A, Zochbauer-Muller S, Sacchi N, Fukuyama Y, Bryant D, Maitra A, Heda S, Fong KM, Thunnissen F, Minna JD and Gazdar AF. (2000). J. Natl. Cancer Inst., 92, 1303–1307. | Article | PubMed | ChemPort |
83. Wang XF, Lin HY, Ng-Eaton E, Downward J, Lodish HF and Weinberg RA. (1991). Cell, 67, 797–805. | Article | PubMed | ISI | ChemPort |
84. Wrana JL and Attisano L. (2000). Cytokine Growth Factor Rev., 11, 5–13. | Article | PubMed | ISI | ChemPort |
85. Yanagisawa J, Yanagi Y, Masuhiro Y, Suzawa M, Watanabe M, Kashiwagi K, Toriyabe T, Kawabata M, Miyazono H and Kato S. (1999). Science, 283, 1317–1321. | Article | PubMed | ISI | ChemPort |
86. Yang X, Letterio JJ, Lechleider RJ, Chen L, Hayman R, Gu H, Roberts AB and Deng C. (1999). EMBO J, 18, 1280–1291. | Article | PubMed | ISI | ChemPort |
87. Zawel L, Dai J, Buckhaults P, Zhou S, Kinzler K, Vogelstein B and Kern S. (1998). Mol. Cell, 1, 611–617. | Article | PubMed | ISI | ChemPort |
88. Zhang Y, Feng X-H and Derynck R. (1998). Nature, 394, 909–913. | Article | PubMed | ISI | ChemPort |
89. Zhang Y, Feng X-H, Wu R-Y and Derynck R. (1996). Nature, 383, 168–172. | Article | PubMed | ISI | ChemPort |
90. Zhu Y, Richardson JA, Parada LF and Graff JM. (1998). Cell, 94, 703–714. | Article | PubMed | ISI | ChemPort |

Acknowledgements

This work is in part supported by a Howard Temin Award (1K01CA78592) and a Burroughs Wellcome Career Award for Biomedical Research (#1676). XH is a recipient of a Howard Temin Award and a Burroughs Wellcome Career Award for Biomedical Research. We thank the following scientists for their generosity in sharing various reagents: Dr S Minucci at European Institute of Oncology for pSG5-RAR, pSG5-PML-RAR, and pcDNA3-PLZF-RAR constructs, Dr D Chakravarti at the University of Pennsylvania for pCMX-RAR and pCMX-RXR constructs, Dr X Liu at the University of Colorado at Boulder for the probe of p15^INK4b, and Dr G Nolan at Stanford University for the E-NX ecotropic packaging cell line. We thank Drs Craig Thompson, Gary Koretzky, Steve Reiner, Mitch Lazar, and Debu Chakravarti for reading the manuscript.