We have cloned a gene, BDP, encoding a protein with homology to the retinoblastoma-binding proteins Rbp1 and Rbp2. It also has homology to DNA-binding proteins such as Bright, a B-cell-specific trans-activator, and the Drosophila melanogaster dead ringer gene product. Like MyoD, Bdp binds to the COOH-terminal region of pRb through its conserved region and to hypophosphorylated pRb. It also binds to the MAR of the immunoglobulin heavy-chain locus. Thus Bdp may contribute to the transcriptional regulation of genes involved in differentiation and tissue-specific expression.
The retinoblastoma gene product (pRb) is a tumor suppressor (1) and is thought to regulate the G1-S transition of the cell cycle (2). It is a nuclear phosphoprotein (3) and is a major target of DNA tumor viruses oncoproteins. Cellular transformation occurs when such oncoproteins as the adenovirus E1A, the SV40 large T antigen, and the human papillomavirus 16 (HPV 16) E7 are able to target and to inactivate functions of the retinoblastoma protein (2, 4, 5).
Binding of the viral oncoproteins requires amino acid residues in a segment of pRb, commonly referred as the ‘pocket region.’ Although the pocket region is necessary for interacting with viral oncoprotein products, it is not sufficient to inhibit cellular proliferation (6). Recently, the COOH terminus of pRb termed the ‘C pocket‘ has also been found to play an essential role in mediating pRb’s functional activity. The COOH-terminal region in collaboration with the pocket region is also essential for the growth-suppressive activity of pRb (6).
The tyrosine kinase c-Abl and the Mdm2 oncoprotein bind specifically to the COOH-terminal region of pRb in vivo (7, 8). The tyrosine kinase activity of nuclear c-Abl is regulated in the cell cycle through this specific interaction with pRb (7). Additionally, a domain in the COOH-terminus of pRb has been shown to bind to the ATP-binding domain of the tyrosine kinase c-Abl, which results in kinase inhibition (7). This interaction demonstrates a link between the growth-inhibitory and growth-stimulatory pathways. The cellular oncoprotein Mdm2 binds and down-regulates p53 function (9). Moreover, Mdm2 binds to the C-pocket of pRb and inhibits pRb growth-regulatory function (8).
We have cloned a novel gene, that we termed BDP, from a cDNA library screening using an EST3 on chromosome 1p31, believed to be the site of a potential tumor suppressor gene (10). Sequence analysis of Bdp revealed homology with two previously characterized pRb-binding proteins, Rbp1 and Rbp2 (11). In vitro binding assays demonstrated that, like Rbp1 and Rbp2, Bdp also interacts with the retinoblastoma protein, although it does not contain the LXCXE pRb-binding motif. The binding site in pRb was found to be the COOH-terminal region. The pRb-binding domain of Bdp, on the other hand, is a highly conserved region that was previously designated as a DNA-binding region of mouse B-cell-specific transcription factor, Bright (12), and a DNA-binding protein encoded from the Drosophila dead ringer (Dri; Ref. 13).
Bdp seems to belong to this family of new DNA-binding proteins because of the high amino acid similarity in the conserved region. Indeed, Bdp binds to the same DNA fragments in the immunoglobulin heavy-chain MAR that Bright binds as assessed by the mobility shift assay. This highly conserved region may function not only as a DNA-binding domain but also as a pRb-binding domain, like the basic helix-loop-helix domain of MyoD (13).
MATERIALS AND METHODS
Isolation and Sequencing of BDP cDNA Clone.
An uncharacterized human cDNA sequence, previously registered in GenBank (nucleotides 74–257, yi78d06.s1), was used as a probe to screen a human testis lgt22 cDNA library. This sequence corresponds to nucleotides 3894 to 4111 of the BDP cDNA sequence. Screening was carried out following standard protocols. 5′-rapid amplification of cDNA ends was performed from human placenta cDNA using a Marathon cDNA amplification kit (Clontech, CA) to retrieve the upstream 62 nucleotide sequences. Sequencing of the recombinant clones was carried out by automated DNA sequencing using the dideoxy terminator reaction chemistry for sequence analysis on Applied Biosystems model 377 DNA sequencers.
The BDP cDNA-containing entire open reading frame was cloned into pBluescript (Stratagene, CA), pGEX-4T (Pharmacia, IL), and pcDNA3.1/His containing Anti-Xpress epitope tag (Invitrogen, CA) from lgt22A. CMV-RB has been described (14). The entire open reading frame of RB (BamHI and StuI fragment) was subcloned into pBluescript. GST-RB (768–928), (835–928), and (768–928;del) were kind gifts of Dr. Y. J. Wang (University of California),and other constructs of the truncated forms of GST-fusion proteins were obtained by enzymatic digestion or by PCR amplification using pBluescript-BDP or CMV-RB. The GAL4 DNA-binding domain containing vectors pAS-RB (379–792) and pAS-RB (793–928) were constructed by cloning PCR-generated fragments into pAS2-1 (Clontech). The GAL4 activation domain containing vector pACT-cyclin D3 was obtained by the screening of human placenta MATCHMAKER cDNA library (Clontech) using pAS-RB (793–928) as a target vector according to the manufacturer’s guidelines. This vector contained the entire open reading frame (nucleotides 81–1962) of cyclin D3.
Mutagenesis was done using a U.SE mutagenesis kit (Pharmacia). The following primers were used: 5′-GTTTGGCCATGATGTGGATTCGGTTGATGG-3′ for GST-BDP (1–56; 240H), 5′-GGTGATCTCCCTCGAGATCTTCTTGTT- 3′ for GST-BDP (1–560; 271S).
Northern Blot Analysis.
The expression pattern of Bdp was verified by detecting mRNA using three commercially available membranes—MTN, MTN II, and Cancer Cell Lines (Clontech). Each membrane was hybridized with 32P-labeled BDP cDNA in ExpressHyb hybridization solution (Clontech) and was washed according to the manufacturer’s instructions. Blots were rehybridized with a 2-kb fragment of human β-actin DNA as a control probe.
Antibody Production and Immunoblot Analysis.
Polyclonal antisera was produced by immunizing a New Zealand White rabbit (HRP) with a bacterially expressed GST-Bdp fusion protein. Immunoglobulin fractions were enriched using EZ-SEP (Pharmacia) and were preadsorbed with acetone powders made from GST-expressed E, coli cells for eliminating GST and bacterial protein-recognizing antibodies. A commercially available polyclonal antibody, C15 (Santa Cruz, CA), was used to detect pRb.
Total cell extracts were prepared by direct lysis of cell pellets with lysis buffer [50 mm Tris-HCl (pH 7.4), 1 mm EDTA, 250 mm NaCl, 0.1% Triton X-100, 50 mm sodium fluoride, 0.1 mm sodium orthovanadate, 0.5 mm DTT, and 1 mm phenylmethylsulfonyl fluoride). Protein concentrations were measured using a SDS-compatible DC protein assay kit (Bio-Rad, CA). Extracted proteins were separated electrophoretically on polyacrylamide gels in the presence of SDS (SDS-PAGE). Proteins within the gels were blotted onto a nitrocellulose membrane electrophoretically.
Immunodetection was performed using the enhanced chemiluminescence (ECL) system (Amersham, IL) according to the manufacturer’s instructions.
Cell Culture, Transfection, and Immunocytochemical Staining.
HeLa (human cervical cancer cell line) and T98G (human glioblastoma cell line) were grown in DMEM with 10% FCS (Life Technologies, Inc., NY). The vector pcDNA3.1/His containing the entire open reading frame of the BDP gene flanked by the Anti-Xpress epitope tag sequence, was lipo-transfected into T98G cells using the Dosper liposomal transfection reagent (Boehringer Mannheim, IN). Transfected cells were selected in 500 μg/ml Genecitin (Life Technologies, Inc.) for 2 weeks. The pool population was incubated in chambers on a slide glass and fixed with 3.5% paraformaldehyde containing 0.1% Triton X-100. To stain the intracellular Bdp, a mouse monoclonal antibody against the Anti-Xpress epitope tag, Anti-Xpress antibody (Invitrogen), and mouse immunoperoxidase staining kit (Santa Cruz) were used.
In Vitro Binding Assays.
GST-fusion proteins were expressed and purified as described previously (15). In vitro translation was performed using a TNT-coupled reticulocyte lysate system (Promega, WI) according to the manufacturer’s guidelines. A mixture of 25 μl of glutathione-agarose beads (Pharmacia) containing 3–5 μg of the respective GST-fusion proteins were incubated with 10 μl of the [35S]methionine-labeled proteins for 90 min at 4°C. Pellets were then washed five times with NETN buffer, and bound proteins were analyzed by SDS-PAGE followed by autoradiography.
This binding was confirmed by incubating HeLa extracts with GST-Bdp fusion protein, followed by immunoblotting with anti-pRb antibody. Exponentially growing HeLa cells (∼80% confluent) were lysed, and the protein was quantified as described above. Protein (800 μg) was diluted in 1 ml of NETN buffer and was incubated with 3 μg of the GST-Bdp fusion protein for 60 min at 4°C. The samples were pelleted and then washed four times in NETN buffer. After separating the proteins on a 7% denaturing SDS-polyacrylamide gel, the samples were then analyzed by immunoblotting as described above using C-15, a polyclonal purified antibody against pRb (Santa Cruz). Exposed films were developed at 15 s and at 30 min.
Mobility Shift Assay.
The mobility shift assay was performed as described previously 12). One ng of 32P-end-labeled DNA probe per lane and 10 ng of in vitro translated Bdp or translation mix from pBluescript vector alone were incubated at 37°C for 15 min. Unbound competitive DNA fragments were preincubated at 37°C for 5 min. After the binding reaction, polyclonal Bdp antibody was added and incubated on ice for 1 h. The probes bf150 and TX125 that we used for mobility shift assays have been described previously (12). They were prepared by PCR amplification and cloned into the pUC19 vector (Stratagene, CA).
BDP Gene Maps to Human Chromosome 15q24 not 1p31.
The EST sequence used to perform our cDNA library screening was obtained from a region on chromosome 1p31 that is thought to harbor a potential tumor suppressor gene on the basis of loss of heterozygosity studies (10). A full-length clone of 4214 nucleotides that revealed an open reading frame of 560 amino acids (Fig. 1) was obtained from a human testis lgt22 cDNA library. Sequence analysis identified two CAG repeats, a CA repeat and a poly(A) signal. No other remarkable motifs were present. However, the sequence on chromosome 1p31 was found to be a pseudogene because it contained no introns. Furthermore, we cloned some intron-containing sequences from a human genomic library, and the true locus of this gene was mapped by fluorescence in situ hybridization to human chromosome 15q24 (data not shown).
Bdp Is Homologous to pRb-binding Proteins and DNA-binding Proteins in a Conserved Region.
The isolated BDP cDNA shares high identity in the conserved region with two previously characterized members of a novel family of genes encoding DNA-binding proteins: Bright, encoding a B-cell-specific trans-activator and the Drosophila melanogaster dead ringer gene product (Dri), (89 and 83%, respectively; Refs. 12, 13; Fig. 2). Several other known proteins share a remarkable degree of homology with Bdp in a more restricted region (amino acids 217–301): (a) the retinoblastoma-binding proteins, Rbp1 (39% identical) and Rbp2 (21% identical; Ref. 11); (b) the X-linked human gene product that escapes X-inactivation (32% identical; Ref. 16; (c) the modulator regulatory factors Mrf1 and Mrf2 (both 30% identical)4; (d) the yeast SWI1 gene, which encodes a member of the SWI regulatory complex (24% identical; 17, 18); and (e) jumonji, a murine gene that is required for neural tube formation (21% identical; Ref. 19). The amino acid sequences of Bdp, Bright, Dri, Rbp1, and Rbp2 are compared in Fig. 2, and their shared areas of identity are highlighted.
Bdp Is a Nuclear Protein with a Distinct Tissue Distribution and with a Predicted Molecular Weight of Mr 61,000.
Northern blot analysis of BDP transcripts showed a dominant species of 4.3 kb highly expressed in placenta, testis, and leukocyte tissue (Fig. 3,A). Human tumor cell lines, K562 and SW480, also expressed abundant BDP mRNA (Fig. 3,B). BDP mRNA was also weakly detected in several other tissues and cell lines (Fig. 3).
Transfection studies followed by immunocytochemical staining confirmed that Bdp is a nuclear protein. The T98G glioblastoma cells were transfected by a pcDNA3.1/His vector containing the entire open reading frame of the BDP gene flanked by a Anti-Xpress epitope tag. Using a monoclonal antibody against the Anti-Xpress epitope tag, we were able to localize staining BDP to the nuclei (data not shown).
Bdp has a predicted molecular weight (Mr) 61,000, and, by Western blot analysis, our polyclonal antibody recognized endogenous Bdp protein in the K562 erythrocytic leukemia cell line and its in vitro translated product with an apparent molecular weight 61,000 as well (Fig. 4).
Bdp Binds to the COOH-terminal Region of pRb through its Conserved Region in Vitro.
Bdp was found to be similar to Rbp1 and Rbp2 in the conserved region (Fig. 2). However, it lacked the LXCXE pRb-binding motif found in Rbp1 and Rbp2. Thus, we wondered whether regions other than the LXCXE motif of Rbp1 may also interact with pRb because even adding 100 mm of Rbp1 peptide containing the pRb-binding motif could not inhibit the Rbp-pRb binding completely (20). It is reasonable to think that Bdp can interact with pRb through the conserved region shared with Rbp1 and Rbp2.
We examined whether Bdp is able to bind pRb in vitro. In vitro translated full-length BDP was able to bind GST-Rb that retained the ‘pocket region’ and the COOH-terminal region (Fig. 5,A, Lane 3). However, it was not able to interact with truncated forms of GST-Rb containing only the A-domain and spacer, or the B domain of the pocket region (Fig. 5,A, Lanes 4 and 5). Binding occurred only in the presence of the COOH-terminal region of GST-Rb (Fig. 5 B, Lane 6).
Furthermore, we examined which amino acids in the COOH-terminal region of pRb specifically interact with Bdp (Fig. 5,B). Bdp was found to bind GST-Rb (768–928; Fig. 5,B, Lane 2) as well as GST-Rb (793–928; Fig. 5,A, Lane 6). However, binding did not occur when the truncated form of either GST-Rb (768–834) or GST-Rb (835–928) were incubated with in vitro translated full-length Bdp (Fig. 5,B, Lanes 3 and 4). A deletion mutant of the COOH-terminal region of pRb (768–928) that does not contain amino acids 774–777 and 836–839 was not able to elicit significant binding (Fig. 5 B, Lane 5). The intact COOH-terminal region of pRb is required for stable binding to Bdp.
To determine the localization of a putative pRb-binding site on Bdp, truncated forms of GST-Bdp were examined for pRb binding (Fig. 6,A). GST-Bdps containing the conserved region were able to bind in vitro translated pRb (Fig. 6,A, Lanes 3 and 5), whereas GST-Bdps that did not contain the conserved region were not (Fig. 6,A, Lanes 4 and 6). Two point-mutant forms of GST-Bdp in which an amino acid was changed at a conserved residue showed reduced binding activity (Fig. 6,B, Lanes 3 and 4) compared with the wild-type form (Fig. 6 B, Lane 2). Binding between the conserved region of Bdp and pRb was specific, which suggests that the three-dimensional organization of this region is important for stable binding.
Bdp Binds Hypophosphorylated pRb.
The phosphorylation state of pRb varies during cell cycle progression, and the growth inhibitory activity of pRb is regulated in this manner. Unphosphorylated and hypophosphorylated pRb, found in resting G1 cells, are believed to be the active forms of the protein needed for growth suppression (20). To date, all of the known pRb-binding proteins interact exclusively with hypo- or unphosphorylated forms of pRb. We examined whether or not Bdp preferentially binds to a specific phosphoform of pRb by in vitro binding assays (Fig. 7). Two mobility shifted bands of the phosphorylated pRb and a nonshifted band from the hypophosphorylated pRb were seen after a short exposure (Fig. 7, Lane 3, left panel). As expected, in vitro translated pRb showed the same mobility as the hypophosphorylated form of pRb (Fig. 7, Lane 4). pRb bound to GST-Bdp also electrophoresed at the same location as the hypophosphorylated pRb after a 30-min exposure (Fig. 7, Lane 2, right panel), which indicated that Bdp preferentially binds the unphosphorylated form of pRb in vitro.
Bdp Is a MAR-binding Protein.
Bdp shares a high identity (89%) with Bright in the conserved region, which is thought to be a DNA-binding region. Bright binds to the DNA fragments in the MAR of the immunoglobulin heavy-chain gene that are located upstream of the S107 variable region promoter, at −424 to −574 (bf150) and −25 to −51 (TX125) from the transcriptional start site (12). We investigated whether Bdp can bind to these DNA fragments, which Bright recognizes (Fig. 8). We found that in vitro translated Bdp was able to bind both bf150 and TX125 fragments (Fig. 8, Lanes 2 and 7), and mobility shifted bands disappeared by adding a 100-fold excess of cold competitor (Fig. 8, Lanes 3 and 8). Adding preimmune serum did not affect these complexes (Fig. 8, Lanes 4 and 9). The addition of polyclonal anti-Bdp antibody caused supershifts (Fig. 8, Lanes 5 and 10); therefore, Bdp binds specifically to the bf150 and the TX125 DNA fragments.
During our search for a tumor suppressor gene related to breast cancer on chromosome 1p31 (10), we cloned a gene encoding a pRb-binding protein, Bdp, by screening a human testis library with an EST sequence as a probe. However, BDP is not a likely candidate for a tumor suppressor gene because the true locus of this gene is on chromosome 15q24 and not on chromosome 1p31.
Sequence analysis confirmed that Bdp shared the highest degree of homology with two members of a previously identified family of DNA-binding proteins, Bright and Dri in the conserved region. Bright is a mouse B-cell-specific trans-activator and binds to the MARs in the immunoglobulin heavy-chain gene locus (12). The DRI gene, dead ringer, was isolated by library screening using the DNA-binding site of the Drosophila engrailed gene encoding a homeodomain protein (13). The high identity shared by these three proteins lies in the conserved region, which suggests that Bdp is a closely related member of a sequence-specific DNA-binding protein family. Another probable addition to this newly recognized protein family has been identified recently as a Homo sapiens DNA-binding protein homologue, named DRIL1 (84% of identity; in GenBank, HSU88047).5 Sequence analysis suggests that DRIL1 is most likely the human homologue of Bright or is very closely related.
The distribution in normal tissues of Bdp was found to be similar to that of Bright (12). Testis and leukocytes displayed the highest level of transcripts. In hematopoietic cancer cell lines, a high level of mRNA expression was detected in the K562 erythrocytic leukemia cell line, whereas a low level of mRNA was detected in the Raji B cell leukemia cell line.
In vitro binding data showed that Bdp binds to the COOH-terminal region of pRb and its binding site, which consists of a highly conserved region thought to be a DNA-binding region also. The binding between Bdp and pRb is specific because mutant forms of either protein disrupt or remarkably reduce binding. Furthermore, the binding between GST-Bdp and in vitro translated pRb was not disrupted in the presence of 250 mm DTT (data not shown), which suggests a strong interaction between these two molecules in vitro. It is not presently known whether Rbp1 and/or Rbp2 bind to the COOH-terminal region of pRb through their conserved region shared with Bdp. However, the presence of the LXCXE-binding motif doesn’t necessarily mean that the interaction with pRb occurs through the pocket region of pRb exclusively. In fact, even if cyclin D2 and D3 contain the LXCXE motif, they are able to interact with the COOH-terminal region of pRb. We confirmed an interaction between in vitro translated cyclin D2 and D3 with the GST-fused COOH-terminal region of pRb (data not shown). In addition, in a yeast two-hybrid system, a stronger interaction was detected between the COOH-terminal region of pRb and cyclin D3 than between the pocket region of pRb and cyclin D3 (data not shown).
The COOH-terminal region of pRb has been reported to interact with the tyrosine kinase c-Abl, the Mdm2 oncoprotein, the transcription factor E2F, and cyclin D (6, 7, 8, 21, 22), but the possible effects of Bdp on the function of these proteins remain to be determined. Our data showed that Bdp interacts preferentially with the hypophosphorylated form of pRb in vitro. Hypophosphorylated pRb has been shown to associate with the nuclear matrix during the G1 phase of the cell cycle and to colocalize with the nuclear matrix protein lamin A and C (23). The COOH-terminal region of pRb may be important for the association with the nuclear matrix in the G1 phase of the cell cycle because it has been shown that Lamin A binds the COOH-terminus (611–928) of pRb in vitro (23). The MAR-binding protein Bdp is also capable of binding the COOH-terminal region of pRb. The possibility that the Bdp-MAR complex associates with the pRb-nuclear matrix complex through the Bdp-pRb interaction is worth examining.
Mobility shift assays showed Bdp specifically bound to bf150 and TX125 DNA fragments in the MAR of the immunoglobulin heavy-chain gene that Bright binds. Several MAR-binding proteins in the intron of the immunoglobulin heavy-chain gene have been reported (12, 24, 25, 26), but each protein is structurally different, and no common regions for DNA binding have been yet identified. Bdp has shown a very high homology in its conserved region with Bright and Dri. We also found other proteins (Rbp1, Rbp2, Xe169, Mrf1, Mrf2, Swi1, and Jumonji) that have a lower degree of homology with Bdp. Moreover, Bright, Dri, Mrf, and Bdp have been shown to elicit a DNA-binding activity. Among all of the these homologous proteins, only Bright has been shown to have MAR-binding properties. Because of the high homology between Bright, Dri, and Bdp and the ability of Bright and Bdp to bind MARs, we think that they could all belong to a family of MAR-binding proteins.
The role of Bdp in the expression of the immunoglobulin heavy-chain gene remains unknown. Bdp expression in B cells by Western blot analysis is much lower than that of Bright expression. Additionally, others have reported that the mobility shifted Bright-bf150 DNA fragment band was supershifted by adding anti-Bright serum with no intensity remaining at the original band location, which meant that protein bound to the bf150 DNA fragment in B cells consisted entirely of Bright (12). It is possible, however, that a Bdp-DNA-binding activity occurs in B cells in certain circumstances. On the other hand, Bdp may be the protein that binds these DNA fragments in non-B cells. It is also possible that Bdp binds other MARs and regulates the expression of other genes. A similar DNA-binding protein SATB1 binds to the regulatory element of γ-globin gene sequences and the regulatory region of the CD8a gene in addition to the MARs in the heavy-chain gene and has been suggested to influence the expression of these genes (27, 28). We are currently examining the MAR-binding activity of BDP with regard to several other genes.
Presently, it is not known whether other homologous DNA-binding proteins in addition to Rbp1 and Rbp2 are able to bind to pRb. In fact, up to now no homologous proteins to Rbp1 and Rbp2, including Bdp, have been shown to modulate the activity of pRb. The binding specificity of Bdp to the hypophosphorylated form of pRb, demonstrated in this report, suggests that the Bdp-pRb complex may be found in the G0-G1 phase of the cell cycle. We are currently investigating the role of our newly cloned BDP in cell cycle regulation and in the possible involvement with pRb in its tumor suppressive activity. On the other hand, it is known that the expression of pRb is required for the myogenic activities of MyoD (29). MyoD binds to the COOH-terminal half of pRb in its basic helix-loop-helix domain that is required for DNA binding and dimerization (29). pRb may regulate myogenic activities through this interaction with MyoD. The presence of a conserved region in our newly cloned BDP for both DNA and pRb binding resembles the basic helix-loop-helix domain of MyoD. In this respect, it may also be possible that the pRb interaction with Bdp and/or other members of this newly identified family contributes to the regulation of the transcriptional activation of various genes involved in differentiation and tissue-specific expression.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by NIH Grants CA39880, CA51083, CA21124, and T32CA09678 (to C. M. C.), and by RO1 CA 60999-01A1, PO1 NS 36466, and PO1 CA 56309.
The abbreviations used are: EST, expressed cDNA sequence tag; MAR, matrix-associating region; NETN buffer, 20 mm Tris-HCl (pH 8.0), 1 mm EDTA, 100 mm NaCl, 0.5% NP40, 1 mm phenylmethylsulfonyl fluoride, and 10 mg/ml leupeptin.
R. H. Whitson, T. H. Huang, B. W. Merrills, T. Asai, and K. Itakura. Manuscript in preparation.
R. D. Kortschak, R. B. Saint, and D. E. Jenne. Submitted for publication.