Resistance to Apoptosis in CTLL-2 Cells Overexpressing B-Myb Is Associated with B-Myb-dependent bcl-2 Induction¹

B-Myb is a widely expressed transcription factor that, together with c-Myb and A-Myb, belongs to the Myb family of transcriptional regulators. All of the family members possess three functional domains: an almost identical NH₂-terminal DNA-binding domain, recognizing the consensus sequence PyAACG/TG; a central acidic transcription activation domain, the sequence of which is only partially conserved among the family members; and a COOH-terminal negative regulatory domain that functions as a transcription repressor. This function is relieved upon cell cycle-regulated cyclin A/cdk2³- or cyclin E/cdk2-dependent phosphorylation in A- and B-Myb but not in c-Myb (1, 2, 3). Overall, B-Myb is more similar to A-Myb than it is to c-Myb, partly because of the presence of an additional region of homology with A-Myb (B-Myb amino acids 308–332) that is absent in c-Myb (2). In several cellular contexts, c-Myb and B-Myb can activate transcription of the same reporter genes (4, 5, 6), leading to the suggestion that they can regulate the same set or a common set of genes (7). Potential MBSs have been found in the regulatory region of several genes, but to date, only a few have been shown to be regulated in a DNA binding-dependent manner by c-Myb (1). In transient transfection assays, B-Myb enhances the promoter activity of B-myb itself, c-myc, HSP70, and cdc2, but there is no conclusive evidence that the activation is dependent on binding to MBSs (2).

c-Myb and A-Myb are characterized by a highly restricted tissue-specific expression (hematopoietic cells and B-lymphocytes and reproductive tissues, respectively), whereas B-Myb is almost ubiquitously expressed. However, as indicated by the phenotype of c-Myb and A-Myb knockout mice, B-Myb is functionally unable to compensate for loss of expression of its more restricted relatives (1). Despite their different expression patterns, a common feature of Myb proteins is their expression in proliferating cells. When overexpressed, they can reduce the growth factor requirements of transfected cells, stimulate cell cycle progression (1), and even overcome a p53-induced G₁ arrest, as shown for B-Myb (8). Moreover, BALB/c-3T3 fibroblasts stably transfected with B-Myb exhibit a partially transformed phenotype and grow efficiently in soft agar (9). In addition to the long-studied but still unclear role in cell cycle regulation, Myb proteins also play an important role in differentiation. Indeed, as cells terminally differentiate and withdraw from the cell cycle, the expression of myb genes is switched off (1, 10). Accordingly, c-Myb overexpression prevents differentiation of immature myeloid and erythroid cells (11), whereas B-Myb overexpression in IL-6-treated M1 myeloid leukemic cells blocks these cells prior to the stage of differentiation associated with cessation of cell proliferation and the subsequent onset of apoptosis (12).

Myb proteins have been recently described as regulators of cell survival in myeloid cells and T lymphocytes, and the bcl-2 gene has been identified as a novel c-Myb target that mediates the anti-apoptotic effect of c-Myb (13, 14, 15). Suppression of c-Myb function has been shown to lead to apoptosis, accompanied by down-regulation of bcl-2 (13, 14). Overexpression of c-Myb protects IL-2-dependent CTLL-2 cells from apoptosis induced by cytokine deprivation, via up-regulation of Bcl-2. However, the c-Myb-dependent transactivation of the bcl-2 promoter is independent from its binding to a MBS (15). The potential role of B-Myb in the regulation of apoptosis is still controversial. B-Myb overexpression reportedly blocked apoptosis of IL-6-treated M1 cells (12), but apoptosis was also accelerated when cells were exposed to transforming growth factor-β (16).

To further assess the role of B-Myb in apoptosis, we studied the apoptotic response of CTLL-2 cells stably transfected with B-myb. We report here that B-Myb overexpression renders these cells resistant to apoptosis by inducing Bcl-2 expression via enhanced transcription. At variance with c-Myb, the transactivation of the bcl-2 promoter occurs via direct binding of B-Myb to a MBS.

CTLL-2 murine T cells were purchased from American Type Culture Collection. Cells were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS, 1% l-glutamine, penicillin/streptomycin (100 μg/ml each), and rhIL-2 (50 units/ml). CTLL-2 cells (5 × 10⁶) were washed twice with PBS, resuspended in 0.5 ml of PBS, and electroporated (Gene Pulser; Bio-Rad, Richmond, CA) at 960 μF and 250 V. Cells were then resuspended in 20 ml of RPMI 1640 supplemented with 10% FBS and 50 units/ml rhIL-2. After 48 h, cells were seeded in 96-well plates in G418-containing medium (0.5 mg/ml). Two weeks later, selected clones were expanded and screened for B-Myb expression by Western blot analysis.

T-lymphoblastic lymphoma CCRF-CEM cells were routinely maintained in RPMI 1640 supplemented with 10% heat-inactivated FBS, 1% l-glutamine, and penicillin/streptomycin (100 μg/ml each).

Equal numbers of cells were washed twice with ice-cold PBS and lysed in 0.1 ml of HEPES buffer [10 mm HEPES (pH 7.5), 150 mm NaCl, 10% glycerol (v/v), 1 mm EDTA, and 1 mm DTT] containing 0.5% (v/v) NP40 in the presence of protease inhibitors (1 mm phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 25 μg/ml aprotinin, 100 μg/ml pepstatin, and 1 mm benzamidine). Lysate preparation and Western blotting were performed according to standard procedures (17). The anti B-Myb antibody was kindly provided by Dr. Robert Lewis (University of Nebraska, Omaha, NE); rabbit polyclonal anti-Bcl-2, anti-Bcl-XL, anti-Bad, anti-Bax, anti-HSP90, and anti-actin antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA); anti c-Myb antibody was from Upstate Biotechnology Inc. (Lake Placid, NY).

Total RNA was extracted from 5 × 10⁶ CTLL-2 cells using TriReagent (Molecular Research Center Inc., Cincinnati, OH) according to the manufacturer’s instructions. RNA was electrophoresed (10 μg per lane), blotted onto a nylon membrane (Amersham Pharmacia Biotech, Piscataway, NJ), and hybridized to a ³²P-labeled 0.9-kb EcoRI fragment of the human bcl-2 cDNA, as described previously (15).

Nuclear run-on assays were performed as described by Chan et al. (18) with some modifications. Briefly, cells (5 × 10⁷) were washed twice with ice-cold PBS, lysed in 10 mm Tris-HCl (pH 7.9), 2 mm MgCl₂, 3 mm CaCl₂, 0.3 m sucrose, and 3 mm DTT in the presence of 0.2% NP40. Nuclei were purified by centrifugation (2000 rpm for 5 min at 4°C) onto a 0.75 m sucrose cushion and resuspended in 200 μl of transcription buffer (1 mm MgCl₂, 70 mm KCl, 15% glycerol, and 1.25 mm DTT), containing 0.25 mm ribonucleotides and 0.5 mCi of [α-³²P]UTP. After 20 min at 26°C, newly synthesized RNA was extracted using TriReagent (Molecular Research Center Inc.), isopropanol-precipitated in the presence of 10 μg/ml tRNA, and purified onto a G-50 RNase-free Sephadex spin column (Boehringer Mannheim, Indianapolis, IN). Equal counts of nascent radiolabeled transcripts were hybridized to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH) containing 5 μg each of the indicated fragments, which were spotted using a slot-blot apparatus. Hybridization was carried out for 72 h at 42°C in 50% formamide, 5× SSC, 1× Denhardt’s solution, 50 mm N₂PO₄ buffer, 100 μg/ml salmon sperm DNA, and 0.1% SDS. Membranes were washed in 2× SSC-0.1% SDS (twice at room temperature and once at 37°C for 20 min each), followed by a single wash in 0.1× SSC-0.1% SDS (at 65°C for 20 min). Autoradiography was performed at −70°C for 15 days with intensifying screens. The bcl-2 probe used for nuclear run-on studies was the same used for Northern blot analysis. The 28S rRNA sequence served as positive control (19).

Plasmids were prepared using standard recombinant DNA methods and PCR techniques. pLXSN, pLXSN-B-myb, CMV-cyclin A, and pBcl-2(1.7)CAT and pBcl-2(1.6)CAT have been already described (14, 19). pBcl-2(1.3)CAT was made by digesting pBcl-2(1.6)CAT with BssHII, blunt-ending with Klenow enzyme, and then digesting with PstI. The resulting 1.3-kb fragment was subcloned into the HindIII-blunted/PstI sites of pCAT basic vector (Promega, Madison, WI). All plasmids were automatically sequenced using the Taq Dye Deoxy Terminator Cycle Sequencing kit (Perkin-Elmer, Norwalk, CT).

CCRF-CEM cells (10⁷) were electroporated as described above with 10 μg of reporter plasmid (pBcl-2-CAT constructs) and effector plasmids (pCDNA3-B-myb+CMV-cyclin A) or empty vector (pCDNA3+CMV-cyclin A) at a 1:1:1 molar ratio, plus 1 μg of pDNApol-βgal, which contains the bacterial β-galactosidase gene driven by the DNA polymerase-α promoter. After 48 h, cells were harvested, and proteins were extracted in 0.25 m Tris-HCl (pH 8) by freeze-thawing and normalized for transfection efficiency by β-galactosidase assay, as described by the manufacturer (Promega). Normalization for protein concentration instead of β-galactosidase activity yielded comparable results. Cellular lysates were incubated with [¹⁴C]chloramphenicol (NEN, Life Science Products, Boston, MA) and acetyl-CoA (Sigma Chemical Co., St. Louis, MO), and CAT activity was measured by TLC followed by autoradiography and densitometry, as described previously (15). Densitometric measurements were performed using a personal densitometer (Molecular Dynamics, Sunnyvale, CA)

Nuclear extracts were obtained from parental and B-Myb-transfected CTLL-2 cells as described previously (20). For EMSA, nuclear extracts (5 μg) in binding buffer [40 mm KCl, 15 mm HEPES (pH 7.9), 1 mm EDTA, 0.5 mm DTT, 5% glycerol, and 1 mm MgCl₂] were incubated for 20 min at room temperature with 3 μg of poly(dI·dC) and ³²P-end-labeled BM42WT, a double-stranded oligonucleotide probe (5 × 10⁴ cpm) corresponding to nt +34 to +58 (relative to the transcription initiation site) of the bcl-2 promoter. Binding reaction mixes were electrophoresed in native 5% PAGE at low ionic stringency (0.25× Tris-borate-EDTA). Gels were dried and exposed to X-ray films for autoradiography.

The sequences used as probes or as competitors in EMSA were as follows: BM42WT, 5′-GCCGCCGCCGCAACACCGAGGTGC-3′; BM42MT, 5′-GCCGCCGCCGGAAGACCGAGGTGC-3′; and GFI-1, 5′-CACCACATAAATCACTGCCTATCC-3′.

Nuclear extracts (100 μg) were incubated at 4°C for 2 h with agarose beads-conjugated-protein A-coupled anti-B-Myb antibody or protein G-coupled anti-c-Myb antibody. After centrifugation, supernatants were saved and used (10 μg) for EMSA, whereas immunoprecipitated proteins were washed four times in lysis buffer [10 mm HEPES (pH 7.9), 150 mm NaCl, 1% NP40, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 μg/ml pepstatin A, 10 mm benzamidine, 100 μg/ml phenylmethylsulfonyl fluoride, 0.2 mm orthovanadate, and 1 mm NaF], resolved by SDS-PAGE, and blotted with the anti-B-Myb or the anti-c-Myb antibody.

Upon transfection of murine CTLL-2 cells with the pLXSN empty vector or with pLXSN-B-myb, several G418-resistant clones were isolated and assessed for B-Myb expression by Western blot analysis. In B-myb-transfected cells, the levels of B-Myb protein were markedly increased, whereas in parental cells, the expression of endogenous B-Myb was barely detectable (Fig. 1,A). CTLL-2 cells are strictly dependent on high concentrations of IL-2 for survival. Indeed, these cells rapidly undergo apoptosis upon not only IL-2 deprivation (21, 22, 23, 24) but also when cultured in the presence of suboptimal concentration of rhIL-2 (5 units/ml; Ref. 15). Thus, we compared the IL-2 requirement of B-myb versus vector-transfected CTLL-2 cells (Fig. 1,B). Upon IL-2 deprivation, ≈ 60% of B-myb-transfected cells were still viable after 24 h, whereas 90% of CTLL-2-LXSN cells were already dead (Fig. 1,B, left). Despite the increased resistance to IL-2 deprivation, B-myb did not confer IL-2 independence: indeed, cells underwent apoptosis 48–72 h after IL-2 withdrawal (data not shown). However, B-myb-transfected CTLL-2 cells did not undergo apoptosis when cultured with 5 units/ml rhIL-2 (Fig. 1,B, right). Indeed, >90% of the CTLL-2-B-myb clones 1 and 15 were alive at 48 h. A slightly lower percentage (80%) of viable cells was found in the culture of CTLL-2-B-myb clones 2 and 13, probably reflecting differences in the level of B-Myb expression (compare clones 1 and 15 versus clones 2 and 13 in Fig. 1 A). Despite some difference in the rate of apoptosis, all clones were routinely maintained in the presence of 5 units/ml rhIL-2.

To exclude the possibility that the reduced IL-2 dependence of B-myb-transfected cells was only apparent, i.e., that the lower requirement was due to autocrine IL-2 production, we performed ELISA assays to detect murine IL-2 secretion in the supernatants of cells transfected with pLXSN or pLXSN-B-myb plasmids and cultured at different concentrations of rhIL-2. No autocrine production of IL-2 was detected in cultures of B-myb-transfected CTLL-2 cells maintained in medium supplemented with 5 units of rhIL-2 (data not shown).

Because B-Myb overexpression can either block or accelerate apoptosis of M1 cells in response to different agents (12, 16), the response of CTLL-2-B-myb cells to different inducers of apoptosis was investigated (Fig. 2). CTLL-2 cells transfected either with the empty vector or B-myb were cultured with 5 units/ml rhIL-2 in the presence or absence of 0.5 μg/ml doxorubicin, 20 mm C2-ceramide, or 1 × 10⁻⁶ m dexamethasone. More than 60% of B-Myb overexpressing cells were viable 48 h after addition of doxorubicin or C2-ceramide, whereas almost complete protection was observed when the cells were treated with dexamethasone. By contrast, ≈10% of vector-transfected CTLL-2 were viable after doxorubicin or ceramide treatment, whereas a slightly higher percentage (≈20%) of surviving cells was observed upon dexamethasone treatment.

To assess potential mechanisms associated with the resistance to apoptosis of B-Myb-overexpressing cells, protein levels of different members of Bcl-2 family were measured. Levels of Bcl-X_L and Bad were essentially identical in CTLL-2-LXSN and CTLL-2-B-myb cells, whereas Bax, Bcl-X_S, and BAG were not detected (data not shown). In contrast, Bcl-2 levels were abundant in B-Myb-overexpressing cells (Fig. 3,A) but detectable only after long exposure in vector-transfected cells (data not shown). By Northern blot analysis, two species of bcl-2 mRNA were identified: a major transcript of 7.5 kb was readily detected in B-myb-transfected CTLL-2 cells but was barely detectable in vector-transfected cells (Fig. 3,B); likewise, very low levels of a minor 2.4-kb transcript were present in B-myb-transfected CTLL-2 cells but were undetectable in vector-transfected cells (Fig. 3 B). Because the M_r 26,000 form of Bcl-2 (Bcl2-2α) is encoded by the 7.5-kb transcript, whereas the M_r 22,000 form (Bcl2-2β) is known to be encoded by the 2.4-kb transcript (24), B-Myb overexpression primarily induces the Bcl2-2α form.

Run-on assay performed using nuclei isolated from B-myb- and vector-transfected CTLL-2 cells showed enhanced bcl-2 transcription in CTLL-2-B-myb cells (Fig. 3 C), suggesting that B-Myb induces an increase in bcl-2 expression, at least in part, by a mechanism of transcriptional regulation.

To investigate whether the enhanced bcl-2 transcription in B-myb-transfected cells was due to B-Myb-dependent activation of the bcl-2 promoter, the ability of B-Myb to transactivate reporter constructs consisting of different deletion mutants of the murine bcl-2 5′-flanking region driving the CAT reporter gene was assessed by CAT assay (Fig. 4) in the human CCRF-CEM T cell line. CCRF-CEM T cells, which express very low levels of endogenous B-Myb (data not shown), were preferred over CTLL-2 cells because CCRF-CEM T cells can be transfected with higher efficiency in transient transfection assays. Accordingly, B-myb cDNA was subcloned into pcDNA3 (Invitrogen, Carlsbad, CA), a CMV-based vector that is more suitable for achieving high levels of expression in human cells. B-Myb is a weak transactivator when tested in transient transfection experiments, but upon cotransfection with cyclin A, its transactivation activity is greatly enhanced (25). Indeed, the cyclin A/cdk2-mediated phosphorylation of B-Myb COOH-terminus potentiates its activity by relieving repression exerted through the COOH-terminal negative regulatory domain (26, 27, 28). A severalfold increase in CAT activity was observed when pcDNA3-B-myb and CMV-cyclin A were cotransfected with CAT reporters driven by bcl-2 promoter segments consisting of the entire 5′-untranslated region and bcl-2 5′-flanking sequences of decreasing length (data not shown). A strong transactivation was still observed using pBcl-2(1.7)CAT and pBcl-2(1.6)CAT, which included only 0.3 and 0.2 kb, respectively, of 5′ flanking sequences upstream of the transcription initiation site (Fig. 4). Compared to transfections with pcDNA3+CMV-cyclin A (Fig. 4, B and C, Lane 1), pcDNA3-B-myb+CMV-cyclin A induced a 7-fold increase in CAT activity when cotransfected with pBcl-2(1.7)CAT, in a 1: 1: 1 molar ratio (Fig. 4, panel B and C, Lane 2). CAT levels driven by pBcl-2(1.6)CAT in the presence of pcDNA3-B-myb+CMV-cyclinA (Fig. 4, panels B and C, Lane 4) were ≈30-fold higher than the basal level in cells transfected with pcDNA3+CMV-cyclin A (Fig. 4, B and C, Lane 3). The pcDNA3-B-myb+CMV-cyclin A plasmids did not transactivate the pBcl-2(1.3)CAT reporter construct lacking the entire 5′ flanking sequence upstream of the transcription initiation site and 0.1 kb of the bcl-2 5′-untranslated region. Indeed, compared with transfections with pcDNA3+CMV-cyclinA (Fig. 4, B and C, Lane 5), pcDNA3-B-myb+CMV-cyclin A slightly repressed the promoter activity of pBcl-2(1.3)CAT (Fig. 4, B and C, Lane 6).

The bcl-2 gene has two promoters, P1 and P2, which are 1.5 and 0.2 kb, respectively, upstream of the ATG translation starting codon (29). By computer-based search, we identified a putative MBS (CAACAC) in the fragment from −200 to +100 (relative to the P1 transcription initiation site) that is not present in the pBcl-2(1.3)CAT reporter (Fig. 5,A). To determine whether B-Myb transactivation of the bcl-2 promoter correlated with a specific DNA-protein(s) interaction, we performed EMSA with nuclear extracts from LXSN- and B-myb-transfected CTLL-2 cells (Fig. 5,B). Using a ³²P-labeled double-stranded oligonucleotide that included the putative MBS (BM42WT) as a probe, we detected a retarded complex in nuclear extracts derived from B-myb-transfected cells (Fig. 5,B, Lane 2). The specificity of the binding between the putative MBS and protein(s) present in nuclear extracts from B-Myb-overexpressing cells was assessed by EMSA performed in the presence of various competitors (Fig. 5,B, Lanes 3–5). For example, formation of the retarded complex was abolished when a 100-fold molar excess of cold BM42WT was used as competitor (Lane 3) but not when an unrelated oligomer, GFI-1 (Lane 4), or a mutated form of BM42WT, BM42MT (Lane 5, in which both Cs of the putative MBS were mutated to G; see Fig. 5 A), were added at the same concentration. Moreover, no retarded complex was detected when the BM42MT oligomer was used as probe, thus confirming that formation of the DNA-protein complex required an intact MBS.

Because anti-B-Myb antibodies suitable for EMSA are not available, the nature of the protein(s) bound to the MBS-containing sequence was characterized by an immunodepletion strategy. Nuclear extracts from B-Myb-overexpressing CTLL-2 cells were depleted of B-Myb by use of a specific anti-B-Myb antibody and tested in EMSA with the ³²P-labeled BM42WT oligomer (Fig. 5 C, Lane 7). Pretreatment with the anti-B-Myb antibody abolished the formation of the retarded complex, indicating that the presence of B-Myb is necessary for BM42WT-protein(s) complex formation.

B-Myb-depleted nuclear extracts were obtained by centrifugation after incubation with protein G-coupled-anti-B-Myb. Supernatant and pellet fractions were then tested for B-Myb by Western blot to confirm the effective depletion from the fraction used for EMSA (Fig. 5 D, top).

Because c-Myb is also expressed in CTLL-2 cells and might bind the site recognized by B-Myb, EMSA were also performed using anti-c-Myb-immunodepleted nuclear extracts. As indicated by Western blotting (Fig. 5,D, bottom), c-Myb was absent from the nuclear extracts used for EMSA. However, the formation of the retarded complex was not abolished when c-Myb-immunodepleted extracts were tested for the ability to bind ³²P-labeled BM42WT (Fig. 5 C, Lane 8), providing additional evidence in support of the specificity of the B-Myb-bcl-2 promoter interaction via the MBS at nt +42 to +47.

A role for B-Myb in the regulation of apoptotic processes was indirectly suggested by demonstrating its requirement for the survival of neuroblastoma cells (30). However, the results of subsequent studies using the myeloid leukemia M1 cell line were apparently contradictory (12, 16), raising questions on the authentic function of B-Myb in apoptosis regulation. In this investigation, we assessed the apoptotic susceptibility of IL-2-dependent CTLL-2 cells stably transfected with B-myb. B-Myb-expressing cells underwent apoptosis with a delayed kinetics after IL-2 withdrawal and exhibited reduced growth factor requirements to the extent that B-Myb-expressing cells were maintained in suboptimal concentrations of rhIL-2 (5 units/ml) that did not allow the survival of parental cells. Further support to the notion that B-Myb exerts an antiapoptotic function was provided by the observation that B-Myb-overexpressing cells are resistant not only to growth factor deprivation but also to other apoptotic stimuli, such as doxorubicin, ceramide, and dexamethasone, which induce cell death by activating different apoptotic pathways. Because B-Myb is a transcription factor, it is conceivable it might regulate the expression of antiapoptotic gene(s) acting at an upstream control point in the apoptotic cascade. Indeed, we found that bcl-2 was transcriptionally up-regulated in B-Myb-overexpressing cells. This finding was not completely unexpected because: (a) several MBSs are present in the bcl-2 promoter; (b) other members of the Myb family, namely, c-Myb and v-Myb, transactivate the bcl-2 promoter in T cells and myeloblasts, respectively (13, 14); and (c) B-Myb can activate transcription of the same reporter genes activated by c-Myb (4, 5), at least in some cellular contexts (6, 7). By means of CAT assay, we identified the region responsible for conferring responsiveness to B-Myb immediately downstream of the P1 promoter, between nt −200 and +100 (relative to the first transcription initiation site). Accordingly, a B-Myb-containing complex was detected when the segment from nt +36 to +58 of the P1 promoter, which contains a putative MBS, was used as a probe in EMSA with nuclear extracts from B-myb-CTLL-2 (but not from parental) cells (Fig. 5,B). Interestingly, the region of the bcl-2 promoter responsible for B-Myb responsiveness differs from those previously reported to be involved in c-Myb-mediated transactivation. Indeed, the critical MBS for DNA binding-dependent c-Myb-mediated bcl-2 transactivation has been mapped few nt upstream of the P2 promoter in the murine thymoma EL4 cells (13). Also, v-Myb regulates bcl-2 promoter activity in ts21 E26-transformed avian myeloblasts via binding to the P2 promoter (14). In c-myb-transfected CTLL-2 cells, in which the bcl-2 promoter is regulated by c-Myb via a DNA binding-independent mechanism, the region responsible for conferring c-Myb responsiveness has been mapped in the P1 promoter (15). Moreover, the B-myb-regulated pBcl-2(1.6) CAT reporter was not transactivated by c-Myb (data not shown). Therefore, bcl-2 appears to be regulated via different mechanisms and through different sites by different members of the Myb family of transcription factors. This finding is particularly interesting because an open question is whether different Myb proteins, despite recognizing the same consensus sequence, activate different promoters. The canonical sequence recognized by Myb family members is the hexanucleotide PyAACG/TG, where the PyAAC core is the essential component of the Myb-binding motif (31, 32). Although all Myb proteins recognize the same binding site in vitro, they may have different binding affinities, depending on the identity of nt flanking the hexanucleotide sequence (33). The MBS downstream of the P1 promoter and interacting with B-Myb (CAACAC) is atypical, with 4 of 6 conserved nt (underlined). However, the interaction of the CAACAC-containing double-stranded oligonucleotide with B-Myb is highly specific because the formation of the DNA-protein complex was completely abolished when: (a) both Cs in the core binding site were mutated into G (Fig. 5,B, Lane 6); and (b) B-Myb was immunodepleted from the nuclear extracts used in EMSA (Fig. 5,C, Lane 7). Moreover, c-Myb immunodepletion did not affect the formation of the DNA-protein complex (Fig. 5 C, Lane 8). Thus, this CAACAC motif might represent the core of a previously unrecognized specific B-Myb-binding site. This hypothesis is further supported by the observation that a specific DNA-protein complex was detected also when the same sequence was incubated with nuclear extracts from B-myb-transfected T98G glioblastoma cell line but not with extracts from untransfected cells.⁴ Additional evidence for the importance of the MBS in B-Myb-dependent transactivation of the bcl-2 promoter could have been obtained by CAT assays with a construct containing a mutated CAACAC motif; unfortunately, despite numerous attempts, we failed to generate a pBcl-2(1.6)CAT construct containing the same mutations shown to prevent the interaction with B-Myb in EMSA.

The identification of bcl-2 as a B-Myb target gene regulated in a DNA binding-dependent manner is interesting because there has been limited evidence in support of DNA binding-dependent transcription regulation by B-Myb (2). Thus, it seems likely that additional B-Myb-regulated target genes identified by more general strategies relying on differential gene expression will be also regulated by B-Myb directly.

In summary, this study shows that B-myb, in addition to its well-recognized function in proliferation and differentiation, also plays an important role in cell survival by up-regulating the expression of the antiapoptotic bcl-2 gene. Because B-myb is ubiquitously expressed in tumor cells, it is conceivable that it may play a role in regulating Bcl-2 levels, which are typically enhanced in several tumors not carrying translocations involving the bcl-2 locus (34, 35). Thus, it would be interesting to examine the functional relationship between B-Myb and Bcl-2 expression in tumor cells and to assess whether interfering with this functional linkage might sensitize tumor cells to apoptosis-inducing treatments.

We thank Arturo Sala for the generous gift of plasmids pLXSN-B-myb and CMV-cyclin A and for helpful advice.