Cell Cycle Arrest and Apoptosis Induced by Notch1 in B Cells*

Notch receptors play various roles for cell fate decisions in developing organs, although their functions at the cell level are poorly understood. Recently, we found that Notch1 and its ligand are each expressed in juxtaposed cell compartments in the follicles of the bursa of Fabricius, the central organ for chicken B cell development. To examine the function of Notch1 in B cells, a constitutively active form of chicken Notch1 was expressed in a chicken B cell line, DT40, by a Cre/loxP-mediated inducible expression system. Remarkably, the active Notch1 caused growth suppression of the cells, accompanied by a cell cycle inhibition at the G1 phase and apoptosis. The expression of Hairy1, a gene product up-regulated by the Notch1 signaling, also induced the apoptosis, but no cell cycle inhibition. Thus, Notch1 signaling induces apoptosis of the B cells through Hairy1, and the G1 cell cycle arrest through other pathways. This novel function of Notch1 may account for the recent observations indicating the selective inhibition of early B cell development in mice by Notch1.

Notch proteins are a transmembrane receptor family that is structurally and functionally conserved from worms to humans. Notch was first identified in Drosophila as a gene involved in neuronal cell fate decision, but this family of receptors is now known to regulate the fate decisions of developing cells in various tissues during embryogenesis as well as in postnatal stages (reviewed in Ref. 1). Upon binding to its ligand, Notch protein is proteolytically processed within the transmembrane domain, and its intracellular domain (Notch-IC) 1 is released (2)(3)(4). Notch-IC translocates to the nucleus and acts as a transcriptional activator in cooperation with a DNAbinding protein, like C promoter binding factor-1 (CBF-1, also known as RBP-Jk) in vertebrates (5)(6)(7)(8), Suppressor of Hairless in Drosophila (9), or Lag-1 in Caenorhabditis elegans (10), together termed CSL proteins. Thus, Notch-IC has been used as a ligand-independent constitutively active form to analyze Notch function in vivo and in vitro (11)(12)(13)(14). The Notch-IC-CSL complex up-regulates the transcription of Enhancer of Split gene encoding basic helix-loop-helix transcription factors in Drosophila and its mammalian homologues, Hairy Enhancer of Split (Hes)-1 (15) and Hes-5 (16), through binding to their promoters. Although Hes family proteins are known to be important components of Notch signaling, Hes-independent pathways are also known to exist in several systems (17)(18)(19).
Notch receptors and their ligands are essential for embryogenesis in mice (20 -24). In the hematopoietic system, Notch1 and Notch2 are expressed in CD34 ϩ hematopoietic progenitor cells (25,26) and a ligand of Notch is found in a subset of bone marrow cells (27). Ligand-mediated activation of Notch1 was shown to inhibit granulocyte colony-stimulating factor-induced differentiation of the myeloid progenitor cell line, 32D (27). In the lymphoid system, Notch1 is strongly expressed in CD4 Ϫ CD8 Ϫ immature thymocytes and its ligands (Jagged1/2) are expressed in thymic stromal cells (28,29). By transgenic approach, it has been found that Notch signaling influences the commitment of CD4/CD8 and ␣␤/␥␦ T cell lineages in the thymus (27,30,31). Recently, it has been reported that the induced deletion of Notch1 gene in adult mice resulted in the impairment of early T cell development and ectopic development of B cells in the thymus (32). Conversely, retroviral expression of constitutively active Notch1 in bone marrow progenitors resulted in a block of early B cell development and the ectopic development of immature T cells in the bone marrow (33). Thus, Notch1 signaling appears to promote T cell development as well as to inhibit B cell development of common lymphoid progenitors primarily destined to develop into B cells.
In mice and humans, B cells are continuously generated in the bone marrow throughout life. In chickens, however, the B cell generation is restricted to a relatively short period of life in a lymphoid organ called the bursa of Fabricius (reviewed in Ref. 34). The bursa is composed of about 10 4 follicles, each of which is colonized by a few B cell precursors expressing surface immunoglobulin (Ig) M at 8 -14 embryonic day (35)(36)(37)(38). The precursors proliferate enormously, and their numbers finally reach to 2-5 ϫ 10 5 /follicle. During this process, extensive diversification of Ig is accomplished by Ig gene conversion (39,40), and most of the developing B cells die by apoptosis prior to emigration into the periphery (34,41,42). The bursa disappears within several months after hatching, and the peripheral B cell pool is maintained by self-replenishing after the involution of the bursa.
To understand how B cell development and homeostasis are regulated in the chicken B cell system, we identify molecules specifically expressed in the bursa and study their functions. Recently, we have found that Notch1 is expressed in the most outer layer of each follicle in the bursa. The Notch1-expressing cells were surrounded by the cells expressing the Notch1 ligand, Serrate2, suggesting a possible role for the Notch signaling in B cell development in the bursa. 2 To understand the meaning of their specific expression pattern and a possible interaction of Notch1 and Serrate2 in the bursa, we analyzed a function of Notch1 in a bursa-derived B cell line, DT40, by expressing a constitutively active form of Notch1 through a new inducible expression system. With this approach, we have revealed a direct effect of Notch1 signaling in cells, namely the induction of cell cycle arrest at the G 1 phase accompanied by apoptosis of the B cell line. In addition, we demonstrated that Hairy1, a downstream target of Notch signaling, also induces apoptosis, but not the G 1 cell cycle arrest, of the same cells.

EXPERIMENTAL PROCEDURES
Construction of Expression Vectors-Chicken Notch1ICS gene (corresponding to amino acids 1748 -2142 of rat Notch1) was amplified by PCR with primers, 5Ј-CTCAAGCTTGGCGCAAGCGGCGCAGGGAGC-ATGGCCAGC-3Ј and 5Ј-CGGAATTCTAGACGGCCGGCTTCAGGTTG-CCGATGTAACTG-3Ј, using a partial chicken Notch1 cDNA 2 as a template (Fig. 1A). The PCR product (1.2 kb) was digested with HindIII and EcoRI, cloned into pBlueScript SK ϩ , and verified for its nucleotide sequence. The HindIII-EcoRI fragment of the Notch1ICS gene was inserted into a multiple cloning site of pEGFP-C1 (CLONTECH, Palo Alto, CA) in frame with green fluorescent protein (GFP) sequence to make pEGFP-Notch1ICS or of pAT7neo, which contains the chicken ␤-actin promoter and double T7 tag in place of the cytomegalovirus promoter/enhancer and GFP sequences of pEGFP-C1, to make pAT7-Notch1ICS. A cDNA encoding a C-terminal portion of Notch1 was amplified by PCR with primers based on the reported sequence (43), 5Ј-GCCCTCTCGGGGCCCCCACGCTGTCCCCCCCGC-3Јand 5Ј-CCC-GAATTCACTTGAAGGCCTCGGGGATGTGTCCCAT-3Ј, using an oligo(dT)-primed single-strand cDNA library of the bursa. The PCR product (1.3 kbp) was cloned into pGEM-T-Easy (Promega, Madison, WI) and verified for its sequence, from which an ApaI-EcoRI fragment was excised and inserted into pEGFP-Notch1ICS to make pEGFP-fulllength intracellular region of Notch1 (Notch1ICF). To make pCALNL5-GFP-Notch1ICS, a 2-kb Eco47III-EcoRI fragment from the pEGFP-Notch1ICS was inserted into SwaI-EcoRI sites of pCALNL5, in which a multiple cloning site including SwaI-EcoRI-SacI-KpnI-SmaI sites had been created at a SwaI site of pCALNLw (44). Similarly, an Eco47III-EcoRI fragment (3.3 kbp) from pEGFP-Notch1ICF was inserted into pCALNL5 to make pCALNL5-GFP-Notch1ICF. Chicken Hairy1 cDNA was amplified by PCR with primers based on the reported sequence (45), 5Ј-CCGAATTCTATGCCCGCCGACACGGGCATGGAAAAACCC-A-3Ј, and 5Ј-CCGGATCCCTACCAGGGGCGCCAGACGGCCTCCCTG-CG-3Ј, using the bursa-derived cDNA library. The PCR product (0.8 kbp) was cloned into pGEM-T-Easy (Promega) and verified for its sequence, from which an EcoRI-BamHI fragment was excised and inserted into pEGFP-C1 to make pEGFP-Hairy1. A SacI-BamHI fragment from the pEGFP-Hairy1 was inserted into pBluescript SKϩ (pBS-Hairy1). Then, a SacI-KpnI fragment from the pBS-Hairy1 including Hairy1 cDNA was inserted in pCALNL5-GFP-Notch1ICS in place of Notch1ICS to make a pCALNL5-GFP-Hairy1. pCAG-Puro-MerCreMer, in which a 3.0-kb HindIII fragment from pAN-MerCreMer (46) was inserted into an XhoI site of pCAG-Puro, was provided by Dr. Michinori Kohara. pCAG-Puro was generated by ligating two ScaI-BamHI fragments from pCAGGS (47) and pBabe-puro (48), containing CAG enhancer/promoter/poly(A) site and a puromycin resistance gene, respectively. 3 Construction of Luciferase Reporter Vectors-The promoter region of chicken Hairy1 gene was isolated by screening of Fix II genomic library from liver (Stratagene, La Jolla, CA), and a 0.7-kb fragment containing a CSL-binding motif (GTGGGAA) was subcloned into pBa-sic2 vector (Toyo Ink, Tokyo, Japan) at upstream of luciferase gene, to make pBasic2-Hairy1wt. The chicken Hairy1 genomic DNA sequence has been deposited in GenBank/EMBL/DDBJ nucleotide data base (accession no. AB045236). The same 0.7-kb fragment but with a deletion of the CSL-binding motif (110 base pairs) was also inserted into the pBasic2 to make pBasic2-Hairy1mt as a negative control.
Cell Lines and Transfection-For the analysis of subcellular localization of Notch1-IC, 3 g of either pEGFP-C1 or pEGFP-Notch1ICS was transfected into NIH3T3 cells (1 ϫ 10 5 ) using trans IT-LT1 (Mirus, Madison, WI). On the next day, the localization of their gene products was visually inspected by fluorescence microscopy. NIH3T3 cells were maintained in Dulbecco's modified Eagle's medium (Sigma, Irvine, United Kingdom) containing 10% heat-inactivated fetal calf serum (FCS; Life Technologies, Inc.) and antibiotics (50 units/ml penicillin and 50 g/ml streptomycin, Life Technologies, Inc.). The bursa-derived B lymphoblastoid cell line, DT40, and its subline CL18 were cultured in RPMI 1640 medium (Sigma) containing 10% heat-inactivated FCS (Life Technologies, Inc.), 5 ϫ 10 Ϫ5 M 2-mercaptoethanol (Life Technologies, Inc.), and antibiotics (Life Technologies, Inc.) as above, at 40°C. For transfection, cells were washed with the culture medium, and adjusted to 2 ϫ 10 7 /ml. Cells (1 ϫ 10 7 ) were transferred to a cuvette and pulsed at 975 microfarads and 250 V in the presence of 30 g of pAT7neo, pAT7-Notch1ICS, or pCAG-Puro-MerCreMer. The transfected cells were selected with 1 mg/ml G418 (Wako, Osaka, Japan) or 1 g/ml puromycin (Sigma). Drug-resistant clones were subjected to Western blotting using anti-T7 mouse monoclonal antibody (Novagen, Madison, WI) or anti-Cre rabbit polyclonal antibody (Novagen). MerCreMer-expressing DT40 cells were further transfected with pCALNL5-GFP, pCALNL5-GFP-Notch1ICS, pCALNL5-GFP-Hairy1, or pCALNL5-GFP-Notch1ICF and selected with 1 mg/ml G418 as above. Drug-resistant clones were treated for 12 h with 10 nM of 4-hydroxytamoxifen (OH-TAM; Sigma), an estradiol analog, and the expression of GFP or GFP fusion proteins was addressed by flow cytometry using FACSort (Becton Dickinson, Mountain View, CA) as well as by Western blotting with anti-GFP mouse monoclonal antibody (CLONTECH). The resultant clones with negligible background and the highest induction of the expression of GFP or GFP fusion proteins were used for the study.
Analysis of Cell Growth and Cell Cycle-The stable transfectants of DT40 cells (5 ϫ 10 5 /ml) were seeded into a six-well plate (4 ml/well) and treated for 12 h with 10 nM OH-TAM to induce the expression of GFP or GFP fusion proteins. Then, the number of live cells were counted using trypan blue and adjusted to 5 ϫ 10 5 /ml every 24 h. At the same time, the cells were also stained with propidium iodide (Sigma) and their DNA contents were analyzed by FACSort as described previously (50).
Northern Blot Hybridization-Transcripts of Hairy1 and MerCreMer genes were analyzed by Northern blot hybridization using the following cDNA probes: a HindIII-BamHI fragment (0.8 kbp) from pBS-Hairy1, and a BamHI-ClaI fragment (0.45 kbp) from pCAG-Puro-MerCreMer. These fragments were labeled with [ 32 P]dCTP by Prime IT II (Stratagene) according to the supplier's instruction. Total cellular RNAs (5 g each) extracted with Trizol (Life Technologies, Inc.) were separated on a 1% agarose gel containing 1% formamide in MOPS buffer, and transferred to a nylon membrane (Biodyne; Pall, East Hills, NY). Hybridization was performed as described previously (51).
Western Blot Analysis-Protein samples were prepared by boiling cells in SDS sample buffer. The samples (2.5 ϫ 10 5 cells/lane) were separated on 7% SDS-PAGE and then transferred to nitrocellulose membrane. The membrane was probed with anti-GFP mouse monoclonal antibody (CLONTECH) or anti-Cre rabbit polyclonal antibody (Novagen). An enhanced chemiluminescence kit (Amersham Pharmacia Biotech) was used to detect horseradish peroxidase-labeled secondary antibodies (Zymed Laboratories Inc., San Francisco, CA).

Construction of a Constitutively Active Form of Chicken
Notch1-To examine the effect of Notch1 signaling in chicken B cells, we constructed vectors expressing an intracellular portion of chicken Notch1 (Notch1ICS, Fig. 1A), which corresponds to a constitutively active form of mouse Notch1 reported previously (30). We then tested whether this chicken Notch1ICS is indeed active. First, the Notch1ICS was transiently expressed in NIH3T3 cells as a fusion protein with GFP and its subcellular localization was determined. As shown in Fig. 1B, the GFP-Notch1ICS protein was found exclusively in the nuclei, whereas the control GFP was distributed diffusely in the cell. The same result was obtained with the chicken B cell line, DT40 (data not shown). We next examined the ability of Notch1ICS in DT40 cells to activate the transcription from the promoter of chicken Hairy1 gene, which is one of Hes family genes and has a CSL protein binding motif in its promoter region (GenBank/EMBL/DDBJ accession no. AB045236). In comparison to the GFP control, the GFP-Notch1ICS strongly activated the chicken Hairy1 gene promoter, but not the mutant promoter with a deletion of the CSL protein binding site (Fig. 1C). Notch1ICS with a T7-tag at the N terminus similarly activated the Hairy1 promoter. These results indicate that the 3 M. Kohara and K. Kohara, unpublished results.

Notch1ICS is constitutively active in chicken B cells.
Growth-inhibitory Effect of Notch1ICS in Bursa-derived Blymphoblastoid Cell Lines-To reveal a function of Notch1 in developing B cells in the bursa, we utilized a bursa-derived B-lymphoblastoid cell line, DT40, as a model system. DT40 cells maintain the characteristics of the bursal B cells, namely expression of specific surface markers such as ChB1 (52) and, more significantly, continuous somatic gene conversion at Ig loci (53). DT40 cells express both Notch1 and Serrate2 weakly, but not Hairy1, suggesting that self ligand-receptor interaction, if any, is not enough to evoke intracellular signaling (see Fig. 6). 2 To address the function of Notch1, we transfected DT40, or its surface IgM-negative variant, CL18, with T7tagged Notch1ICS expression vector. Despite repeated transfection experiments, however, we could not establish any stable transfectants from DT40, and only one from CL18 expressing Notch1ICS, which showed very slow proliferation rate (data not shown). Cell cycle analysis of the Notch1ICS-transfectant showed that it contained high proportion of apoptotic cells, compared with the parental or mock-transfected CL18 cells (Fig. 2). Similar effect of the Notch1ICS was observed in a few transfectants obtained from 249L4 cells, another bursa-derived B cell line expressing Notch1 (data not shown). Thus, Notch1ICS expression appears to be lethal for the B cell lines tested above, and occasionally arising transfectants suffer a severe growth disadvantage with continuous apoptosis in a fraction of cells.
Induced Expression of Notch1ICS Causes G 1 Cell Cycle Inhibition and Apoptosis in DT40 Cell Line-To avoid the cell lethality by the constitutive expression of Notch1ICS and to analyze the direct effects of Notch1ICS on the proliferation and/or survival of the cells, we applied a Cre/loxP-mediated inducible system (54) for the expression of Notch1ICS. An expression vector encoding the chimeric Cre recombinase, Mer-CreMer, containing a ligand binding domain of the mouse estrogen receptor at both ends, was first introduced into DT40 cells. A stable transfectant strongly expressing the MerCreMer protein was selected by Western blotting, then transfected with the expression vectors in which GFP or GFP-Notch1ICS fusion genes were separated from the promoter by loxP-flanked neomycin-resistance gene (Fig. 3A). In the resultant transfectants (GFP or GFP-Notch1ICS clones), loxP-mediated recombination and the following expression of GFP or GFP-Notch1ICS genes can be induced by treatment with OH-TAM. Flow cytometric analysis showed that more than 97% of the transfectants expressed either GFP or GFP-Notch1ICS proteins upon treatment with 10 nM OH-TAM for 12 h (Fig. 3B). Two of the GFP-Notch1ICS clones (N#1-1, N#3-1) and a control GFP clone (G#1-13), each showing negligible fluorescence before OH-TAM treatment, were used for further analyses.
Since OH-TAM itself inhibited the growth of DT40 cells slightly (data not shown), transfectants were treated with OH-TAM for the first 12 h to induce the expression of transgenes, then washed and further cultured in normal medium. The GFP-Notch1ICS clones proliferate constantly in the absence of Notch1ICS expression, but the proliferation was gradually inhibited by the induction of Notch1ICS, which first became evident 3 days after the induction (Fig. 4, B and C). The control GFP clone proliferates at a constant rate irrespective of the induction of GFP (Fig. 4A).
To characterize the growth regulation by Notch1ICS, cell cycle profile was analyzed by flow cytometry every 24 h of the culture after the induction of Notch1ICS expression (Fig. 5A). The proportion of apoptotic cells increased gradually and reached over 25% by day 4 after the expression of Notch1ICS (a). The proportion of the cells in the S phase started to decrease at day 2 and became about half of the starting population by day 4 (c), while that in the G 2 /M phase slightly decreased (d). The proportion of the cells in the G 0 /G 1 phase was roughly unchanged during the experimental period (b), indicat- ing a relative increase of this population among live cells. Control cells expressing GFP did not grossly alter their cell cycle profile in these culture periods. Representative cell cycle profiles on day 4 is shown in Fig. 5B. The ratio of the proportions of the cells in the G 0 /G 1 versus S phases significantly increased by expression of the Notch1ICS, but not by expression of the GFP. These results indicate that Notch1 signaling inhibits cell cycle progression at the G 1 phase and induces apoptosis of DT40 cells, although the kinetics of these changes appear to be rather slow compared with that of apoptosis of the same cells induced by surface IgM cross-linking (55).
Induced Expression of Hairy1 Causes Apoptosis in DT40 Cell Line-In mice, Hes1 gene is the best characterized target of Notch signaling and its transcription is up-regulated by Notch-IC-CSL complex bound to its promoter region. We have also demonstrated that Hairy1 gene promoter was transactivated by Notch1ICS in a CSL-binding site-dependent manner (Fig.  1C). Accordingly, the level of endogenous Hairy1 transcript was up-regulated by the induction of Notch1ICS expression in the DT40 cell transfectants (Fig. 6A). Therefore, it is possible to speculate that the Hairy1 protein mediates the Notch1 signaling that induces cell cycle inhibition and/or apoptosis of DT40 cells. To test this possibility, we next generated the DT40 transfectants in which the expression of GFP-Hairy1 fusion protein can be induced by the same vector system as described above (Fig. 3A). The expression of the GFP-Hairy1 fusion mRNA was induced by OH-TAM treatment in the transfectants to almost the same level as the endogenous Hairy1 mRNA induced by Notch1ICS (Fig. 6A). The induction of the GFP-Hairy1 expression resulted in growth suppression as evident already on day 1, and the number of live cells did not increase after day 2 (Fig. 6B). Cell cycle analysis revealed that the proportion of the apoptotic cells obviously increased 2 days after the Hairy1 induction and reached to 45% by day 4 (Fig.  6C). In sharp contrast to the case of Notch1ICS-induction, the proportion of the cells in the G 0 /G 1 phase started to decrease as early as day 1 and continued to decrease until day 4, whereas that in the S/G 2 /M phases did not change markedly by day 2 and later decreased in proportion to the increase of apoptotic cells. Thus, the induction of Hairy1 expression caused earlier and stronger apoptosis compared with that induced by Notch1ICS, although it did not cause G 1 cell cycle inhibition.
We also tried to establish stable transformants by repeated transfections of conventional constitutive expression vectors encoding T7-tagged Hairy1. However, we failed to obtain a single clone from DT40 cells, and obtained several clones from 249L4 cells all of that showed a marked apoptosis in culture and gradual decrease of Hairy1 expression and/or outgrowth of Hairy1-negative variants (data not shown).

C-terminal Parts of Notch1 Are Necessary for Full Activity of Notch1 Intracellular Domain to Induce G 1 Cell Cycle Arrest and
Apoptosis-Above results are consistent with a scenario in which the Notch1ICS complexed with CSL protein up-regulates the expression of Hairy1 gene, and then the Hairy1 protein regulates genes to induce apoptosis, whereas other mediator(s) than Hairy1 being activated by Notch1ICS induces cell cycle arrest at the G 1 phase. However, the Notch1ICS-induced apoptosis and G 1 arrest appeared to be rather modest in terms of the timing of onset and the extent (Fig. 4 and 5). We thought this might be due to the lack of the C-terminal amino acids of Notch1 in the Notch1ICS used above. Therefore, we next constructed an inducible expression vector coding for a full-length intracellular portion of Notch1 (Notch1ICF) as a GFP fusion protein, and induced its expression in DT40 cells in the same system as above (Fig. 3A). The GFP-Notch1ICF localized in the nucleus of the cells and induced the Hairy1 promoter activity as well as the expression of the endogenous Hairy1 gene to the similar level to that GFP-Notch1ICS did (data not shown, Fig.  6A). The expression level of GFP-Notch1ICF protein was less than that of GFP-Notch1ICS protein after induction (Fig. 7A). Nevertheless, the induction of the Notch1ICF resulted in a strong cell cycle arrest at G 1 phase (Fig. 7B). Proportion of cells in the G 1 phase had already started to increase and that in the S/G 2 /M to decrease in the induction period (see day 0), and most of live cells were arrested at the G 1 phase as early as on day 1. The Notch1ICF expression also strongly induced apoptosis as evident by day 2, and the proportion of the apoptotic cells increased thereafter. The G 1 cell cycle arrest and apoptosis induced by the Notch1ICF were much stronger and more rapid than those by the Notch1ICS shown above (Fig. 5). These results indicated that the full length of the Notch1 intracellular domain is required for its full activity to induce apoptosis and G 1 cell cycle arrest, not affecting the level of transactivation of the effects of an active form of chicken Notch1 in a bursaderived B lymphoblastoid cell line, DT40 cells, using a new inducible expression system. Our results indicate that the Notch1 signal induces both apoptosis and G 1 cell cycle arrest in the B cells, whereas Hairy1 only mediates apoptosis. In addition, we have found that the C-terminal region of Notch1, including PEST domain, is necessary for full activity.
A human homologue of Notch, TAN1, was first identified as a proto-oncogene whose intracellular region was translocated into the T cell antigen receptor ␤ locus in T lymphoblastic leukemia/lymphomas (56). The ability of TAN1-IC to induce T cell leukemia was confirmed in mice by its retroviral introduction into mouse bone marrow cells (57). In addition, Notch1-IC was shown to rescue thymoma cells and T cell hybridomas from glucocorticoid-and T cell antigen receptor-mediated apoptosis, respectively (58,59). These observations suggest that Notch1-IC is able to promote the proliferation and survival of cells of the T cell lineage, in sharp contrast to its activities in B cells as described in this paper. Such an opposite function of Notch1 signaling between T and B cell lineages may account for the recent findings that the Notch1 signaling is necessary for the T cell development, whereas it inhibits B cell development in mice (32,33). Thus, shortly after the T/B lineage commitment of lymphoid progenitors, Notch1 signaling may promote proliferation and differentiation of T cell progenitors, whereas inhibit those of B cell progenitors.
Recently, it was shown that Hes1 is necessary for the selfrenewal of multipotent progenitor cells in mouse brain (60) and  Fig. 4. C, before (Ϫ) or after induction of GFP-Hairy1 by OH-TAM, cells were cultured and analyzed for their DNA contents as in Fig. 5. Shown is a representative result of three independent experiments with one stable clone and several experiments with other clones, all of which showed similar results.
for expansion of early T cell precursors in mouse thymus (61), indicating that Hes1 promotes proliferation of these cells. Our results, however, demonstrate that Hairy1, whose expression was up-regulated by the Notch1-IC, induced apoptosis of the B cell lines. Thus, these two proteins may have different effects in controlling apoptosis despite their sequence homology. Alternatively, the function of Hes family proteins may be cell type-specific.
Overexpression of Hairy1 in DT40 cells resulted in earlier and severer apoptosis compared with that caused by Notch1ICS, although the expression level of the GFP-Hairy1 gene in the former was equivalent to that of the endogenous Hairy1 gene induced by the latter. This suggests that Notch1ICS may also signals to attenuate the function of Hairy1, presumably through translational or post-translational modification of Hairy1, or through affecting the downstream of Hairy1 in the pathway leading to apoptosis. On the other hand, Hairy1 did not cause G 1 cell cycle arrest in contrast to Notch1ICS or Notch1ICF, indicating that the Notch1-ICinduced G 1 cell cycle arrest is mediated by Hairy1-independent mechanism(s). Deltex is another Notch-binding protein whose structure and function are conserved from flies to humans. It functions as a positive regulator of Notch signaling (62,63). The Deltex expression is up-regulated by Notch1-IC (58) and Deltex inhibits the transcriptional activity of E47, a basic helixloop-helix protein that is necessary for early B cell development (18,63). In B cells, E47 homodimers activate transcription of several genes that are critical for B cell development, such as IgH, 5, VpreB, and Rag1 (64 -66), as well as B cell-specific activator protein (BSAP)/Pax5 (66,67). BSAP/Pax5 is a transcription factor also necessary for early B cell development (68,69). It was observed that the BSAP/Pax5 gene was strongly expressed in B cell lymphomas through its translocation into Ig heavy chain locus (70,71). In addition, antisense oligonucleotide-mediated suppression of BSAP activity caused growth inhibition of B cells, but not of T lymphoma or plasma cell lines (72). These observations indicate that BSAP/Pax5 promotes proliferation of B cells. Taken together, it seems possible to imagine a scenario in which Notch1 signaling inhibits E47 activity through Deltex, and thus down-regulates BSAP/Pax5 expression, resulting in growth inhibition of B cells. Experiments to test this possibility are in progress with our DT40 cell system.
We demonstrated here that the C-terminal region of Notch1 including the PEST domain strongly enhanced the induction of G 1 cell cycle arrest and apoptosis by Notch1 signaling in DT40 cells. It was previously shown that C-terminal region of mouse Notch1 protein acts as positive regulatory domain for CSL-dependent transcriptional activation and works as transactivation domain by itself (73). Although Hes1 gene expression is known to be activated by CSL, we have observed no significant enhancement of Notch1-IC-induced endogenous Hairy1 expression as well as Hairy1 gene promoter activation by the addition of the C-terminal region (Fig. 6A, data not shown). Therefore, a role of the C-terminal domain in CSL-mediated transcriptional activation may differ among the target genes. To support this, Schroeder and Just (74) reported that presence/absence of the C-terminal domain of mouse Notch1 did not affect the CSL- A, comparison of the expression level of Notch1ICF protein with Notch1ICS protein. The stable transformants were cultured with OH-TAM, and then lysed in SDS sample buffer. Protein samples were subjected to Western blotting probed with anti-GFP antibody (upper). The same filter was reprobed with anti-Cre antibody to verify the amount of loaded protein (lower). B, kinetic analysis for cell cycle of a stable DT40 cell transfectant carrying a Notch1ICF-inducible vector. The cells were treated with OH-TAM, cultured, and analyzed for their DNA contents as in Fig. 6C. Shown is a representative result of three independent experiments with one stable clone and several experiments with other clones, all of which showed similar results. mediated enhancement of granulocyte colony-stimulating factor-induced differentiation of a myeloid progenitor cell line. In our system, it seems that the C-terminal region of chicken Notch1 positively regulated transactivation of certain genes responsible for G 1 cell cycle arrest and apoptosis.
A few B cell precursors undergo extensive proliferation and Ig gene conversion in the bursa to enlarge their diversity. It was estimated, however, that the vast majority of the newly generated B cells die in situ, and that only 5% of the cells generated daily emigrate to the periphery (41,42). In the periphery, mature B cells are resting until they encounter the antigen. As mentioned above, B cells located in the most outer layer of each follicle in the bursa express Notch1, and are surrounded by the cells expressing the Notch1 ligand. 2 Therefore, it is possible to speculate that the B cells are arrested in their proliferation by Notch1 signaling and die by apoptosis unless some survival signal is provided. In this way, Notch1 might contribute to select functional B cells and to promote maturation of B cells before emigration into the periphery.