MRG15 Activates the B-myb Promoter through Formation of a Nuclear Complex with the Retinoblastoma Protein and the Novel Protein PAM14*

The MORF4-RelatedGene on chromosome 15 (MRG15) is a member of a novel family of genes originally identified in studies to reveal cell senescence-inducing factors. MRG15 contains several predicted protein motifs, including a nuclear localization signal, a helix-loop-helix region, a leucine zipper, and a chromodomain. These motifs are commonly associated with transcription factors, suggesting that MRG15 may likewise function as a transcriptional regulator. To examine the potential function(s) of MRG15, we sought to identify cellular factors associated with thisMRG family member. In this regard, we have found that both the retinoblastoma tumor suppressor (Rb) and a novel nuclear protein PAM14 (Protein Associated with MRG,14 kDa) specifically associate with MRG15. We have further demonstrated that these interactions require the helix-loop-helix and leucine zipper domains of MRG15. Interestingly we have found all three proteins present in a multiprotein complex, suggesting that at least some of their functions may be interdependent. Although the functions of PAM14 have yet to be elucidated, Rb has several well characterized activities, including repression of E2F-activated promoters such as that of B-myb. Significantly we have demonstrated that MRG15 blocks the Rb-induced repression of this promoter, leading toB-myb promoter activation. Collectively these results suggest that MRG15 regulates transcription through interactions with a cellular protein complex containing Rb and PAM14.

Replicative senescence, or the terminal loss of proliferative potential exhibited by normal cells in culture, is viewed as both a model for aging at the cellular level and as a tumor suppressor mechanism (1). In studies to identify cell senescence-related genes, we cloned MORF4 1 (MORtality Factor on chromo-some 4), a novel gene that induces a senescent phenotype upon introduction into a subset of immortal human cell lines (2). It was subsequently shown that MORF4 is a member of a novel gene family whose protein products share several common structural motifs, including a nuclear localization signal, a helix-loop-helix region, and a leucine zipper. As these domains are frequently found in transcriptional regulators, we hypothesized that the MORF4 family members may function similarly to regulate transcription. Consistent with this possibility we have previously established that members of this family are localized within the nucleus of cells (2).
Upon cloning each member of the MORF4 gene family, we found that in addition to MORF4, only two of the other family members were expressed, MRG15 and MRGX (MORF4-Related Genes found on chromosomes 15 and X, respectively) (2). MRG15 is of particular interest because it is expressed in a wide variety of human tissue types and is highly conserved across multiple species, including flies (Drosophila melanogaster), worms (Caenorhabditis elegans), yeast (Schizosaccharomyces pombe and Saccharomyces cerevisiae) and plants (Arabidopsis thaliana) (3). This cross-species conservation suggests that MRG15 possesses an activity fundamentally important to one or more cellular processes.
Sequence alignment data demonstrate that MRG15 is nearly identical to MORF4 (96% amino acid similarity). Unlike MORF4, however, MRG15 fails to induce a senescent phenotype upon introduction into immortal cell lines, and MRG15 RNA levels decline with the onset of senescence (2). The most striking structural difference between MRG15 and MORF4 is that MRG15 contains an additional predicted domain at its amino terminus that codes for a region known as a chromodomain. The chromodomain is a motif identified in several proteins that function as negative or positive regulators of transcription, including proteins from D. melanogaster and S. cerevisiae such as the Msl-3, polycomb, HP1, and SWI/SNF proteins (4 -7). These regulators do not appear to bind DNA directly but rather associate indirectly with specific sites on chromatin via interactions with transcriptional repressors or activators as well as with proteins that influence chromatin accessibility to such transcriptional regulators. The Msl-3 protein, for example, has been implicated in the regulation of dosage compensation in Drosophila by acting in a multimeric complex that binds to hundreds of specific sites on the male X chromosome and induces hypertranscription likely through modification of chromatin structure (8). Interestingly MRG15 exhibits sequence similarity to the Drosophila Msl-3 protein over its entirety, suggesting that MRG15 may similarly affect gene expression in cells by associating with specific transcription factors in multimeric nuclear complexes.
The fact that the chromodomain and leucine zipper motifs have been implicated in protein-protein interactions suggested that MRG15 associates with one or more cellular factors. Therefore, identification of such interacting proteins is expected to provide critical insight into MRG15 function. We report here the identification of two MRG15-associated factors as a novel nuclear protein PAM14 (Protein Associated with MRG, 14 kDa) and the retinoblastoma tumor suppressor protein (Rb). Rb is known to repress the promoters of many genes, including those involved in cell cycle progression (9 -12). We elected to study one paradigm of an Rb/E2F repressed promoter, the B-myb promoter (13)(14)(15), and have shown that MRG15 blocks Rb-mediated repression of this promoter. These results suggest that MRG15 forms a nuclear protein complex with PAM14 and Rb that may function to control transcription from Rb-regulated promoters.

EXPERIMENTAL PROCEDURES
Antibodies-Commercially available rabbit hemagglutinin (HA) polyclonal antibodies (Santa Cruz Biotechnology) and rabbit polyclonal Rb (C-15) antibodies (Santa Cruz Biotechnology) were used for immunoblot analysis. The polyclonal MRG15 antibody was generated by immunizing rabbits with a peptide containing the sequence from the 5Ј-chromodomain region of the protein (DEWVPESRVLK) and was used for immunoblot analysis. Commercially available HA monoclonal antibodies (Roche Molecular Biochemicals) or Rb monoclonal antibodies were used for immunoprecipitation assays. Commercially available anti-rabbit and anti-mouse secondary antibodies (Pierce) were also used.
Plasmids-Wild-type and mutant B-myb promoter constructs in the pGL2 luciferase reporter (wt myb and mut myb) were generously provided by N. Dyson (13). The mutant B-myb promoter-reporter construct contains a mutation in the E2F binding site (deletion of nucleotides Ϫ208 to Ϫ206) that eliminates Rb binding and repression. To isolate the full-length PAM14 cDNA, an internal primer was synthesized and used for rapid amplification of cDNA ends with the AP1 primer (CLONTECH). The resulting amplicons were directly ligated to the PCRII vector using a TA cloning strategy (Invitrogen). The inserts corresponding to the 5Ј-end of the PAM14 cDNA were sequenced, and the following primers were used to amplify the open reading frame from a human heart library (CLONTECH): 5Ј-CCGA-AGCTTCCACCATGCGGCCCCTGG-3Ј and 5Ј-GGCGGATCCAGGG-TGTCAGCCAATCTC-3Ј.
Cells, Cell Culture, and Transfections-CMV-transformed human fibroblasts (CMV-MJ-Hel1), EJ bladder carcinoma-derived cells, HeLa cervical carcinoma cells, and Saos2 osteosarcoma cells were obtained from the American Type Culture Collection (Manassas, VA) and maintained in Hanks' minimum essential medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum. All transfections were carried out using LipofectAMINE Plus (Life Technologies, Inc.) according to the manufacturer's instructions. To generate stable clones of cells expressing HA-tagged PAM14, CMV-MJ-Hel1 and EJ cells were transfected with PAM-HA:pcDNA3.1. Twenty-four hours post-transfection, 3,000 -10,000 cells/60-mm dish cells were subcultured into Earl's minimum essential medium (Life Technologies, Inc.) supplemented with G418 (1 mg/ml). After a 2-week incubation at 37°C in a 5% CO 2 incubator, several clones were isolated and maintained in medium containing G418. PAM14-HA expression in each clone was verified by immunoblot analysis.
Microscopy-CMV-MJ-Hel1 cells were seeded at 2 ϫ 10 5 cells/35-mm dish and transfected 24 h later with the indicated plasmids. At 24 h post-transfection the cells were observed under a fluorescence microscope (Olympus). For confocal microscopy, CMV-MJ-Hel1 cells stably expressing PAM14-HA were seeded at 2 ϫ 10 5 cells on glass coverslips in 35-mm tissue culture dishes. Approximately 24 h post-transfection the cells were transfected with the MRG15:GFP construct. At 48 h post-transfection, the cells were then washed with phosphate-buffered saline, fixed with 3% paraformaldehyde, and permeabilized with 0.1% Triton X-100 for 15 min. The cells were incubated first with rabbit HA polyclonal antibodies (1:500) and mouse monoclonal green fluorescent protein (GFP) antibodies (1:500). The cells were then stained with fluorescein isothiocyanate-conjugated goat anti-rabbit IgG secondary antibodies (1:250) (Molecular Probes) and Texas-red-conjugated goat anti-mouse IgG secondary antibodies (1:250) (Molecular Probes). All antibodies were diluted in IF buffer (50 mM Tris-HCl, pH 7.5, 200 mM NaCl) and incubated at 37°C. Coverslips with attached cells were rinsed briefly in a 0.5 g/ml 4Ј,6-diamidino-2-phenylindole solution, and coverslips were affixed to microscope slides with mounting medium (VectorShield). Microscopy was performed using an Applied Precision DeltaVision microscope (Issaquah, WA) fitted with an Olympus IX70 microscope. Images were acquired via wide field sectioning using fluorescent light. The stacked images, usually 20 -25 sections, were subjected to point spread function analysis for better image quality on Silicon Graphics software (SGI, Mountain View, CA). Magnification of all cells was maintained at 600ϫ. PAM14-HA was seen as red, MRG15-GFP as green, 4Ј,6-diamidino-2-phenylindole nuclear staining as blue.
Yeast Two-hybrid Analysis-A yeast two-hybrid screen was performed using the Matchmaker two-hybrid system (CLONTECH). MRG15:pGBT9 (bait) and a human fibroblast cDNA library (a gift from J. Campisi) cloned into the activator plasmid pGAD424 were transformed into the PJ69-4A yeast strain (a gift from J. Campisi), which contains three different inducible reporter markers for adenine, histidine, and ␤-galactosidase synthesis. Plasmid DNA was isolated using DNA extraction buffer (2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM Tris, pH 8.0, 1 mM EDTA), 0.2 ml of phenol/chloroform (1:1), and 0.3 g of glass beads. Plasmid DNA derived from each yeast colony was transformed into the DH5␣ Escherichia coli strain, and the DNA inserts were sequenced to determine the identity of the interacting clones. The prey plasmids identified in the screen were each retransformed with the MRG15 bait plasmid by the lithium acetate method (16) and selected on plates lacking tryptophan and leucine. These yeast were streaked on plates under selection for adenine, tryptophan, and leucine synthesis where growth indicated an interaction between the two proteins that activated an inducible reporter marker. These colonies were subsequently streaked on plates lacking histidine, adenine, tryptophan, and leucine to increase the stringency in the tests for interaction. In addition, liquid ␤-galactosidase reporter assays (Promega) were performed to assess the strength of the protein interactions.
GST Pull-down Assays-The bacterial cell line BL21 (Stratagene) was transformed with the indicated plasmid, and expression of the GST fusion protein was induced by addition of isopropyl-␤-D-galactoside for 3 h at 37°C. A small aliquot was removed from the bacterial cultures, lysed in 2ϫ sample buffer, separated by SDS-PAGE, and Coomassiestained to verify expression of the fusion proteins. The bacterial lysates were subsequently harvested and purified on Sepharose 4B glutathione beads (Amersham Pharmacia Biotech). For GST pull-down assays with cell lysates, EJ or Saos2 cells transiently transfected with the indicated expression plasmids were harvested in Nonidet P-40 lysis buffer (20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 250 mM NaCl, 0.5% Nonidet P-40) supplemented with a protease inhibitor mixture (Calbiochem). Beadimmobilized GST-tagged proteins were incubated with the lysates for 3 h, then washed four times in RIPA buffer (150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 50 mM Tris, pH 8.0), solubilized in 2ϫ sample buffer, run on an SDS-polyacrylamide protein gel, and either transferred to a nitrocellulose membrane (Bio-Rad) for immunoblot analysis or silverstained using the Silver Stain Plus Kit (Bio-Rad) according to the manufacturer's instructions. For GST pull-down assays with in vitro transcribed/translated products, in vitro transcription/translation was performed on the indicated plasmid constructs using the T7 Single Tube Protein kit (Novagen). In vitro transcribed/translated protein products were then inoculated with the GST proteins and glutathione-Sepharose beads in binding buffer (50 mM Tris, pH 8.0) for 2 h, then washed four times with Nonidet P-40 lysis buffer, run on a 12% polyacrylamide gel, dried, and developed by autoradiography.
Immunoprecipitations-Protein A/G BioMag beads (Polysciences) were preincubated with 8 g of HA or Rb monoclonal antibodies for 1 h at 4°C. Nuclear extracts of EJ and CMV-MJ-HeL1 clones stably expressing PAM14-HA were prepared by resuspension of cells in buffer A (25 mM Tris-HCl, pH 7.5, 50 mM KCl, 2 mM MgCl 2 , 1 mM EDTA, 5 mM dithiothreitol) and higher salt extraction of proteins in the nuclear pellet in buffer B (25 mM Tris-HCl, pH 7.5, 0.42 M NaCl, 1.5 mM MgCl 2 , 0.5 mM EDTA, 1 mM dithiothreitol, 25% sucrose). The lysate was then incubated with the antibody-bound beads for 3 h at 4°C, washed four times in Nonidet P-40 lysis buffer, separated by SDS-PAGE, and immunoblotted with the indicated antibodies.
Luciferase Assays-EJ cells were transfected with the indicated expression plasmids and harvested with Reporter Lysis Buffer (Promega), and the luciferase activity was determined using the Luciferase Assay Kit (Promega) and a Monolight 2010 luminometer. The data was normalized to the amount of protein in the samples as determined by the Bradford assay (Bio-Rad). Luciferase assays were performed in triplicate to verify reproducibility, and protein assays were performed in duplicate. Statistical analyses were performed using the Student's t test.

RESULTS
Interaction of MRG15 with a Novel Nuclear Protein, PAM14 -MRG15 encodes several motifs implicated in mediating protein-protein interactions, suggesting that this protein associates with other cellular factors. Therefore, to identify potential MRG15-interacting proteins, we screened a human fibroblast cDNA library by yeast two-hybrid assay using fulllength MRG15 as bait. A total of 2 ϫ 10 5 transformants were screened, yielding seven colonies that grew under selection. The strength of each interaction was assessed by ␤-galactosidase reporter activity upon co-transfection of MRG15 and the plasmids expressing each interacting protein into yeast (data not shown), and the strongest interacting protein, PAM14, was isolated for further analysis. As further verification of this interaction, yeast co-expressing PAM14 and MRG15 or the positive controls SV40 T antigen and p53 were competent for growth under Ϫadenine/tryptophan/leucine (ϪATL) and Ϫhistidine/adenine/tryptophan/leucine (ϪHATL) amino acid selection, whereas yeast co-expressing the SV40 T antigen and either PAM14 or MRG15 failed to grow under selection (Table  I). Thus, MRG15 and PAM14 are not intrinsic trans-activating proteins. Interestingly, unlike many leucine zipper proteins, MRG15 was not found to homodimerize in this system (Table I).
Cloning and sequencing of the full-length PAM14 revealed it to be a 1205-nucleotide cDNA encoding a novel 14-kDa protein of 127 amino acids (GenBank TM accession number AF116272) (Fig. 1). Secondary structure analysis (Expasy) predicts this protein primarily consists of a coiled-coil helical structure throughout its length. To physically examine the MRG15-PAM14 interaction, we performed GST pull-down assays and found that the MRG15-GST fusion protein but not GST alone bound HA-tagged PAM14 (PAM14-HA) ( Fig. 2A). Likewise, in the converse experiment, PAM14-GST but not GST alone bound HA-tagged MRG15 (MRG15-HA) ( Fig. 2A). Furthermore, the interaction between these two cellular factors is direct as MRG15-GST retained the ability to bind in vitro transcribed and translated PAM14 (Fig. 2B).
To determine the domain(s) of the MRG15 protein required for interaction with PAM14, we generated a panel of MRG15-GST deletion mutants lacking the predicted structural domains (Fig. 2C) and demonstrated that the wild-type and mutant GST fusion proteins were expressed to equivalent levels (Fig. 2D). In GST pull-down assays, the MRG15 mutant lacking the amino-terminal chromodomain (ϪCHR) bound PAM14 with similar affinity as wild-type MRG15 (Fig. 2E). By contrast, the MRG15 mutants lacking either the helix-loop-helix region (-HLH) or leucine zipper (ϪLEU) failed to bind PAM14 (Fig. 2E). Thus, these results suggest that the helix-loop-helix and leucine zipper domains but not the chromodomain of MRG15 are important for interaction with PAM14. We next sought to determine the subcellular localization of PAM14 by examining cells expressing GFP-tagged PAM14 using fluorescence microscopy. Interestingly GFP-PAM14 localized to the nucleus of human cells but was excluded from the nucleoli (Fig. 3A), an expression pattern similar to that previously observed for GFP-tagged MRG15. In accordance with these results, we observed co-localization of PAM14 and MRG15 in HeLa cell nuclei by confocal microscopy (Fig. 3B).
MRG15 Interacts with Rb-Although we predict the interaction between MRG15 and PAM14 to be significant, the lack of knowledge regarding PAM14 function made it difficult to gain insight into MRG15 function based on its association with this novel protein. We therefore attempted to identify additional MRG15-associated proteins with more established cellular functions. To this end, we performed GST pull-down assays of EJ cell nuclear extracts using MRG15-GST to identify MRG15associated cellular factors. MRG15-GST-bound proteins were resolved by SDS-PAGE and visualized by silver staining, and bands unique to the MRG15-GST pull-down assay not detected in pull-down assays with GST alone were catalogued. As an additional control, the MRG15-GST protein not incubated with lysate was examined simultaneously to identify contaminating bacterial protein bands and MRG15-GST breakdown products. Pull-down assays were performed multiple times under conditions of increasing stringency, and in each of these experiments we consistently observed a MRG15-associated protein with an PAM14 ϩ ϩ ϩ a Growth on yeast plates lacking tryptophan and leucine (ϪTL) indicates the presence of both bait and prey plasmids in the yeast strain, whereas growth on plates also lacking adenine (ϪATL) or adenine and histidine (ϪHATL) indicates interaction between the bait and prey proteins as this leads to activation of a Gal4 promoter and subsequent transcription of genes required for adenine and histidine synthesis. b Gal4 DBD represents the bait plasmid containing the indicated gene as a fusion with the Gal4 DNA binding domain as well as a gene allowing selection on media lacking tryptophan. c Gal4 TD represents the prey plasmid containing the indicated gene as a fusion with the Gal4 transactivation domain as well as a gene allowing selection on media lacking leucine. approximate molecular mass of 100 kDa (p100) (Fig. 4A). To identify this protein, we obtained antibodies to several known nuclear proteins of the appropriate molecular weight, including Rb (17), p107 (18), and p130 (18) and tested them in immuno-blots of MRG15-GST pull-down reactions. Using this screening approach, we succeeded in identifying Rb as a second MRG15associated cellular factor (Fig. 4B). To confirm the identity of Rb as a MRG15-associated protein, we performed similar GST pull-down experiments using lysates from Saos2 cells, which are known to lack functional Rb (Fig. 4B). As expected, we failed to detect the 100-kDa MRG15-associated band in these experiments, further suggesting that this MRG15-interacting protein is Rb.
To determine the region(s) of the MRG15 protein required for the Rb interaction, we utilized the MRG15-GST deletion mutants described above (see Fig. 2C) in GST pull-down assays. Similar to our results with PAM14, the MRG15 deletion mutant lacking the amino-terminal chromodomain bound Rb with similar strength as wild-type MRG15, whereas the MRG15 mutants lacking the helix-loop-helix or leucine zipper domains failed to bind Rb (Fig. 4C). Thus, the helix-loop-helix and leucine zipper regions are important both for the interaction of MRG15 with PAM14 as well as with Rb.
MRG15, PAM14, and Rb Are All Present in a Multiprotein Complex-Our observations that similar MRG15 domains are important for binding both PAM14 and Rb might suggest that these interactions are mutually exclusive. Therefore, to determine whether all three proteins co-exist in a multiprotein complex or whether MRG15 binds only one of the cellular factors at a time, we performed immunoprecipitations from EJ and CMV-MJ-HeL1 cell lines stably expressing PAM14-HA. Upon immunoprecipitation of PAM14-HA from nuclear extracts of these cells, we observed co-immunoprecipitation of both MRG15 and Rb, suggesting that all three proteins were present in a multiprotein complex (Fig. 5, A and B). Similar results were obtained when nuclear extracts from EJ and CMV-MJ-Hel1 clones were immunoprecipitated with an Rb antibody (Fig. 5, A and B), further suggesting that MRG15 is able to simultaneously complex with PAM14 and Rb.
MRG15 Relieves Rb-mediated Transcriptional Repression-We next wanted to examine the functional significance of the MRG15-Rb interaction. Rb has well established roles in proliferation and differentiation through its ability to regulate gene transcription (19). However, Rb is not a classical transcription factor in that it does not possess sequence-specific DNA binding activities. Rather Rb is recruited to promoters through interactions with transcription factors, such as those of the E2F family. In the case of E2F, Rb functions as a potent negative regulator of transcription in that it complexes with E2F family members to block transcription from E2F-responsive promoters (17). The association of Rb with MRG15 led us to hypothesize that MRG15 might influence the transcriptional regulatory activities of Rb. We therefore examined the effect of MRG15 on the well characterized E2F-responsive B-myb promoter. During G 1 growth arrest, transcription of the B-myb gene is repressed upon sequestration of E2F into inactive complexes by Rb (13)(14)(15). To examine the effect of MRG15 on Rb-mediated repression of the B-myb promoter, EJ cells were co-transfected with a luciferase reporter construct driven by the B-myb promoter and increasing amounts of a MRG15 plasmid. We observed a statistically significant increase (3-5-fold, p Ͻ 0.01) in luciferase activity in a dose-responsive manner in

FIG. 4. MRG15 binds Rb in vitro.
A, MRG15-GST specifically binds a 100-kDa cellular factor. EJ cell nuclear extracts (NE) were subjected to GST pull-down assays with the indicated GST fusion protein and then silver-stained to identify MRG15-associated bands. B, MRG15 associates with the 100-kDa Rb tumor suppressor protein in EJ cells but not in Saos2 cells lacking Rb. EJ or Saos2 cell nuclear extracts were subjected to GST pull-down assays with the indicated GST fusion protein and then immunoblotted with Rb antibodies. the presence of 0.5 and 1.0 g of MRG15 (Fig. 6A). This effect was specific to wild-type MRG15 as a MRG15 deletion mutant lacking the leucine zipper motif failed to stimulate B-myb promoter activity (Fig. 6A). To verify that the observed increase in B-myb promoter activity was due to a counteraction of Rbmediated repression rather than a general MRG15 transcriptional activation function (i.e. providing nonspecific access to histone acetylases) we performed similar luciferase reporter assays with a mutant B-myb reporter construct lacking E2F binding sites. Although this mutant has higher basal B-myb promoter activity, this activity was not further increased upon introduction of MRG15 (Fig. 6B, p Ͼ 0.04), suggesting that MRG15 is not a general transcriptional activator. In addition, we failed to detect activation of the B-myb promoter by MRG15 in RbϪ/Ϫ Saos2 cells (Fig. 6C), further suggesting that the activation of the B-myb promoter by MRG15 is mediated through an inhibition of transcriptional repression by Rb. As expected, the levels of wild-type and mutant B-myb promoter activity in Saos2 cells are similar (Fig. 6C), presumably due to a lack of Rb-mediated repression of the wild-type B-myb promoter. Collectively, these observations suggest that the interaction of MRG15 with Rb can result in the inhibition of Rbmediated B-myb transcriptional repression. DISCUSSION The recent discovery of the cell senescence-inducing protein MORF4 has lead to the subsequent identification of several MORF4-related genes, although gene expression has only been verified for two other members of this family, MRG15 and MRGX. Interestingly, although both MRG15 and MRGX exhibit a high degree of sequence similarity to MORF4, neither of these genes exhibit the senescence-inducing activity characterized for MORF4. However, the possibility remains that these MRG family members function in related aspects of cellular proliferation. Consistent with this notion is the observation that members of this family possess several putative protein interaction motifs commonly found in proteins with roles in transcriptional regulation. MRG15 represents a particularly interesting member of this family, both because it is widely expressed, even during embryonic development, and because it is conserved across multiple species including vertebrates, insects, yeast, and plants (2). Therefore, to gain functional insight into MRG15, we examined interactions of this protein with additional cellular factors. We report here the identification of two MRG15-associated cellular proteins, the retinoblastoma tumor suppressor Rb and the novel protein PAM14. Significantly we demonstrate that all three proteins exist in a multiprotein complex in cells and, furthermore, that MRG15 can relieve Rb-mediated transcriptional repression. From these data, we propose a model whereby the interaction of MRG15 with Rb and possibly PAM14 negatively regulates Rb-induced E2F-responsive promoter repression and, as a result, may facilitate cell cycle progression.
The related proteins Rb, p107, and p130 are all members of the pocket protein family of transcriptional repressors named for their pocket-like protein interaction domain that typically binds LXCXE motifs of target proteins. We found that MRG15 specifically interacts with Rb but fails to bind either p107 or p130. This observation, coupled with the fact that MRG15 lacks the canonical LXCXE motif, suggests that the interaction between MRG15 and Rb does not occur through the pocket domain common to this family of proteins. The fact that PAM14 is predicted to possess extensive helical structure and that several domains of Rb are similarly predicted to form helices, suggests that MRG15 may target helical domains of these proteins. Both the helix-loop-helix and leucine zipper domains of MRG15, which are important for PAM14 and Rb binding, are similarly predicted to have a helical structure (3). It is of note that many transcription factors, such as the B/Zip proteins, likewise contain coiled-coil domains that mediate specific interaction with their partners (20).
Although Rb is represented in the human fibroblast cDNA library analyzed in the yeast two-hybrid screen, 2 we failed to detect its interaction with MRG15 in this system. One possible explanation is that more transformants need to be analyzed to detect the Rb association in this assay. Alternatively the interaction between MRG15 and Rb may be mediated through another protein(s) not present in the library. Furthermore, MRG15 and/or Rb may require specific post-translational modifications, which may not occur in yeast, to promote their interaction. In this regard, both MRG15 and Rb are known to contain multiple phosphorylation sites that may regulate specific protein interactions. In particular, Rb is phosphorylated and thereby inactivated by multiple cyclin-dependent kinases during cell cycle progression (21). One possibility is that Rb phosphorylation may promote its interaction with MRG15, perhaps as an additional negative regulator of Rb function during the cell cycle.
Our observation that MRG15 de-represses the E2F-responsive B-myb promoter suggests that MRG15 may function to antagonize the transcriptional repressive functions of Rb. Although the specific E2F binding site(s) on Rb have yet to be elucidated, the fact that E2F lacks the LXCXE motif suggests that the interaction does not involve the Rb pocket domain (22). As MRG15 likewise lacks the LXCXE motif, one possibility is that it shares a similar Rb binding site with E2F. In this scenario, the interaction of MRG15 with Rb could be envisioned to displace E2F from Rb, thereby facilitating activation of E2Fresponsive promoters. With regard to possible mechanisms of MRG15 function, the fact that this protein was not a self-activator in the yeast two-hybrid assay but was able to activate the B-myb promoter may suggest that the transcriptional activation functions of MRG15 are not direct but rather mediated through interactions with specific cellular factors. This idea is further substantiated by both the failure of MRG15 to activate either a mutant B-myb promoter lacking E2F binding sites and by the inability of the MRG15 leucine zipper deletion mutant, which is impaired for Rb and PAM14 binding, to de-repress the wild-type B-myb promoter. Furthermore, the failure of MRG15 to activate the wild-type B-myb promoter in RbϪ/Ϫ Saos2 cells suggests that the transcriptional activation functions of MRG15 arise via its disruption of Rb function. Our data indicate that MRG15, PAM14, and Rb are all present within a multiprotein complex in cells, perhaps indicating that the interactions of MRG15 with both of these factors are required for MRG15 functions. However, the multitude of MRG15-associated proteins observed in GST pull-down assays by silver staining (data not shown) suggests that there are likely several additional cellular factors in MRG15-containing complexes, one or more of which may be required for the transcriptional de-repression activities of MRG15. Collectively these data support the hypothesis that the interactions of MRG15 with multiple cellular factors lead to de-repression of the B-myb promoter as well as possibly other E2F-responsive elements.
The ability of MRG15 to increase B-myb promoter activation has implications for a potential role of MRG15 in cell cycle progression. The B-myb protein is an established transcriptional activator believed to promote cell cycle progression through stimulation of multiple effectors of cell growth, including Cyclin D1 and cdc2 (9,23). Furthermore, overexpression of B-myb has been demonstrated to promote bypass of p53-induced G 1 arrest (10), suggesting that B-myb also induces cell growth through inhibition of growth inhibitory molecules. Bmyb expression correlates with cell cycle progression in that it is repressed during G 0 and early G 1 phases but stimulated in late G 1 and S phases. Significantly this mRNA expression pattern is similar to that of MRG15; we have observed an approximately 2-fold increase in MRG15 mRNA levels at 4 -8 h post-serum stimulation of normal quiescent fibroblasts, and these elevated mRNA levels persist up to 28 h poststimulation FIG. 6. Activation of the E2F-responsive B-myb promoter upon overexpression of MRG15. A, wild-type but not mutant MRG15 stimulates transcription from the B-myb promoter. EJ cells were transfected with 0.5 g of a luciferase-expressing plasmid under the control of the B-myb promoter either alone or together with 0.25-1.0 g of a plasmid expressing wild-type MRG15 or 1.0 g of a MRG15 deletion mutant lacking the leucine zipper motif. Cells were lysed 24 h post-transfection, and the amount of luciferase activity in each lysate was determined using a luminometer and normalized according to the amount of protein in each sample. The data are compiled from three independent experiments, each performed in triplicate. B, MRG15 fails to activate a mutant B-myb promoter lacking E2F-responsive sites. EJ cells were transfected with 0.5 g of a luciferase-expressing plasmid under the control of either the wild-type B-myb (wt myb) promoter or a mutant B-myb (mut myb) promoter lacking the E2F-responsive sites either alone or together with 1.0 g of a plasmid expressing MRG15, and luciferase activity in each transfection was determined as described in A. The data are compiled from three independent experiments, each performed in triplicate. C, MRG15 fails to activate the wild-type B-myb promoter in RbϪ/Ϫ Saos2 cells. Saos2 cells were transfected with 0.5 g of a luciferase-expressing plasmid under the control of either the wild-type B-myb promoter or a mutant B-myb promoter lacking the E2F-responsive sites either alone or together with 0.25-1.0 g of a plasmid expressing wild-type MRG15. Luciferase activity in each transfection was determined as described in A. The data are from one representative experiment, which was performed in triplicate.
(2). Young fibroblast cells 4 -8 h poststimulation are in the early to mid-G 1 phase of the cell cycle and enter S phase at 16 h postexposure to serum with maximal DNA synthesis at 24 h poststimulation. Thus, like B-myb, MRG15 expression correlates with cell cycle progression, and it therefore represents a putative growth-stimulatory factor. It is tempting to speculate that such potential proliferative functions of MRG15 are due to its ability to antagonize one or more activities of Rb, resulting in enhanced B-myb expression. Interestingly the PAM14 sequence has been deposited in the data base as PGR1 (Gen-Bank TM accession number AF116272), a protein induced in quiescent lymphocytes following stimulation to enter the cell cycle. 3 Thus, the ability to interact with PAM14 may also play an important role in the putative growth-stimulatory activities of MRG15.
The facts that MRG15 is expressed in a wide variety of tissues and is conserved across a large number of species suggest an essential function for this cellular factor. Our data showing that MRG15 forms multiprotein complexes within the nucleus that may antagonize Rb function further support this contention. Future studies aimed at examining the effect of MRG15 on additional Rb/E2F-repressed promoters and identifying additional cellular factors present within MRG15 complexes are therefore expected to provide critical insight into the role of this protein in the cell cycle and possibly other processes.