Evidence That C Promoter-binding Factor 1 Binding Is Required for Notch-1-mediated Repression of Activator Protein-1*

Cell fate determination in invertebrate and vertebrate systems is regulated by the Notch signaling pathway. Four mammalian Notch genes, Notch 1–4, encode differentially expressed transmembrane receptors. The canonical Notch pathway involves proteolytic liberation of the Notch-1 intracellular domain (NIC-1), which activates CSL (CBF1, Su(H), and Lag-1)-mediated transactivation. We showed previously that NIC-1 also represses activator protein-1 (AP-1)-mediated transactivation. The N-terminal RAM (RBP-Jκ associated molecule) domain of NIC-1 was required for both activation and repression. To investigate the mechanism of AP-1 repression, we tested whether distinct sequences within the RAM domain mediate activation versus repression. We analyzed the capacity of RAM domain mutants to bind endogenous CBF1, to activate CSL-mediated transactivation, and to repress AP-1. A mutant lacking 20 amino acids of the RAM domain (Δ1759–1778) resembled the RAM domain deletion mutant in being defective in all activities. Analysis of 14 deletion and alanine substitution mutants revealed a correlation between CBF1 binding, CSL-mediated transactivation, and AP-1 repression. Stably transfected K562 cells could only tolerate very low level expression of wild-type NIC-1 and NIC-1 mutants retaining activation/repression activities. By contrast, transcriptionally compromised NIC-1 mutants accumulated at high levels. These results support a model in which the binding of NIC-1 to CBF1 is required for AP-1 repression and reveal a powerful cell-sensing mechanism that suppresses the levels of transcriptionally competent NIC-1.

The Notch signaling pathway is a major system for controlling cell fate during the development of organisms as diverse as insects, nematodes, and mammals (1,2). Although this pathway was originally identified and studied in Drosophila, genes homologous to components of the pathway have been cloned from numerous metazoan organisms. Vertebrates have four paralogs of Notch, Notch-1-4, and six genes encoding Notch ligands, Jagged-1 and -2, Delta-1, and Delta-like-1, -3, and -4 (3)(4)(5)(6). The canonical Notch pathway involves activation of Notch through the binding of a transmembrane ligand on a neighboring cell. Ligand binding induces consecutive proteo-lytic cleavages catalyzed by metalloprotease tumor necrosis factor ␣-converting enzyme (7) and ␥-secretase (8), which release the intracellular domain of Notch (NIC), 1 allowing it to undergo nuclear translocation. In the nucleus, NIC regulates target gene transcription by binding to DNA-bound CSL (CBF1, Su(H), and Lag-1) proteins (9), thereby converting CSL from a repressor into an activator (10,11). A host of coregulators, including mastermind (12,13), CBP/p300 (14,15), SMRT (16), and Ski-interacting protein (17), have been implicated in the control of NIC-mediated transactivation.
Given the diverse macromolecular interactions in which Notch engages and its crucial biological functions, not surprisingly, the domain organization of Notch is highly conserved (18). The Notch extracellular domain contains up to 36 tandem epidermal growth factor-like repeats and three cysteine rich Lin-12/Notch repeats, which function in ligand binding and Notch activation. The N terminus of NIC contains the RAM domain that physically associates with CSL proteins, which bind preferentially to the DNA sequence CGTGGGAA (9,19) and variations thereof (19). The region C-terminal to the RAM domain, containing six cdc10/Ankyrin repeats, mediates additional protein interactions. The cdc10/Ankyrin repeats also interact weakly with CSL (20). The region C-terminal to the cdc10/Ankyrin domain has been implicated in various protein interactions and is important for transactivation (TAD) (21), although the PEST motif contributes to the control of Notch stability (22). Drosophila Notch and mammalian Notch-1, -2, and -3 contain nuclear localization signals (23) and an OPA sequence of unknown function (24).
We reported that the human Notch-1 intracellular domain (NIC-1) not only activates CSL-mediated transactivation but also represses AP-1-mediated transactivation (25). Because endogenous AP-1 target genes were repressed and similar concentrations of NIC-1 mediated activation versus repression, repression appeared to be physiologically relevant. Repression correlated with nuclear localization of NIC-1 and required the RAM domain. Repression was not accompanied by altered c-Jun N-terminal kinase-dependent signaling events, which activate AP-1. Based on the crucial roles of Notch in developmental processes such as hematopoiesis (26 -30) and vasculogenesis (31)(32)(33)(34) and the diverse regulatory functions of AP-1, including roles in hematopoiesis and vasculogenesis (35)(36)(37)(38), Notch-AP-1 cross-talk is likely to be highly significant.
Since our discovery of Notch-AP-1 cross-talk, two additional reports described similar interactions (39,40). NIC-1 repressed * 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  transcription of E6/E7 papillomavirus genes, and this was accompanied by up-regulation of Fra-1, a component of inhibitory AP-1 complexes, and down-regulation of c-Fos, a component of activating AP-1 complexes (39). However, Fra-1 and c-Fos (25) expression were unaffected by NIC-1 in K562 cells, under conditions in which AP-1-mediated transactivation was repressed. Furthermore, NIC-1 expression in K562 cells did not impact upon AP-1 DNA binding activity in nuclear extracts in vitro, indicating that NIC-1 did not induce the assembly of stable repressive heterodimers. Another member of the AP-1/ATF superfamily of basic leucine zipper transcription factors, B-ATF, is up-regulated by NIC-1 in Epstein-Barr virus-negative B-cell lymphoma BJAB cells (40). B-ATF is expressed highest in hematopoietic tissues and assembles into a repressive complex with c-Jun. Because B-ATF is not expressed in HeLa cells, NIC-1 represses AP-1 in these cells (25), and as noted above, NIC-1 does not affect AP-1 DNA binding activity in vitro, it is unlikely that B-ATF is a general mediator of inhibitory NIC-1-AP-1 cross-talk. Thus, although Fra-1 and B-ATF mediate NIC-1-AP-1 cross-talk in certain contexts, we expect that multiple modes of cross-talk exist, analogous to the inhibitory steroid receptor-AP-1 cross-talk, which has been studied for more than a decade without a unifying mechanism (41,42).
Here, we investigated the mechanism of how Notch represses AP-1 and describe the identification of eight amino acids within the RAM domain that are crucial for CSL binding, activation, and repression. This analysis also revealed a powerful cellsensing mechanism for selectively down-regulating NIC-1, which can be overridden by substituting only four amino acid residues with alanines.
Alanine substitution mutagenesis of NIC-1 was accomplished by insertion of oligonucleotide fragments containing alanine substitutions into the NIC-1 expression vector after digestion with appropriate restriction enzymes. For NIC-1(1759 -1762A), NIC-1(1763-1766A), and NIC-1(1767-1770A), the pBabe-NIC-1 expression vector was digested with BamHI and SalI. Purified NIC-1(BamHI-SalI) was subjected to another digestion with Bsu36I. Purified pBabe(SalI-BamHI) and NIC-1(Bsu36I-SalI) were ligated with oligonucleotides containing Ala substitutions. NIC-1(1771-1774A) and NIC-1(1775-1778A) were generated by sequential high fidelity PCR reactions. A short fragment from the NIC-1 BamHI site to the sequences that were substituted with Ala residues and a long fragment from the same sequences to the NIC-1 SacII site were generated by PCR using two primer sets. This strategy was designed such that the two fragments had substantial overlapping sequences surrounding the Ala substitution region. The fragments were then used as templates to generate the NIC-1(BamHI-SacII) fragment containing Ala substitutions using the forward primer for the short fragment and the reverse primer for the long fragment. After digestion with BamHI and SacII, this fragment was ligated with pBabe(SalI-BamHI) and NIC-1(SacII-SalI).
Cell Culture-The human erythroleukemia cell line K562 was maintained as described previously (25,45). The cells were propagated in Iscove's modified Eagle's medium (Biofluids) containing 10% fetal bovine serum and 1% penicillin/streptomycin (Invitrogen) (complete IMEM) in a humidified incubator at 37°C in the presence of 5% carbon dioxide.
Transient Transfections-K562 cells (5 ϫ 10 5 ) were collected by centrifugation at 240 ϫ g for 5 min at 4°C and resuspended in 4 ml of complete IMEM. Plasmid DNAs (1 g of reporter and 2 g of effector) were suspended in 150 l of IMEM, incubated with Superfect (4 l/1 g DNA; Qiagen) for 10 min at room temperature, and then added to the cells. The cells were incubated for 26 h post-transfection and then treated with TPA (final concentration, 5 nM) or the vehicle (Me 2 SO). After incubating for 16 h, the cells were harvested, and the cell lysates were assayed for luciferase activity. The luciferase activity was normalized by the protein content of the lysates, determined by Bradford assay using ␥-globulin as a standard.
To determine the expression levels of wild-type NIC-1 and NIC-1 mutants, DNA was transiently transfected into K562 cells using DM-RIE-C (Invitrogen). DNA (24 g) and DMRIE-C reagent (72 l) were diluted in 2.0 and 3.0 ml of OPTI-MEM I medium (Invitrogen), respectively. The two solutions were combined and incubated at room temperature for 45 min. K562 cells (9 ϫ 10 6 ) were collected by centrifugation at 240 ϫ g for 5 min at 4°C, washed once with ice-cold phosphate-buffered saline, and resuspended in 1.0 ml of OPTI-MEM I medium. The cells were then added to the mixture of transfection components. After incubation at 37°C in a CO 2 incubator for 4 h, complete IMEM (12 ml) containing 15% fetal bovine serum was added to the cells. The expression levels of wild-type NIC-1 and NIC-1 mutants were determined by Western blotting 40 h post-transfection.
Stable Transfection-K562 cells were stably transfected using Lipo-fectAMINE 2000 (Invitrogen). K562 cells (1 ϫ 10 6 ) were collected by centrifugation at 240 ϫ g for 5 min at 4°C, washed once with ice-cold phosphate-buffered saline, and resuspended in 800 l of OPTI-MEM I medium. Plasmid DNA (2 g) was added to 100 l of OPTI-MEM I medium, incubated with LipofectAMINE 2000 (4.5 l/g of DNA in 100 l of OPTI-MEM I medium) for 30 min at room temperature, and then added to the cells. After incubating for 5 h at 37°C in a 5% CO 2 incubator, complete IMEM (2 ml) containing 15% fetal bovine serum was added to the cells. Puromycin (1.5 g/ml) (Sigma) was added 72 h post-transfection to select for stable transfectants, and expression levels were measured by Western blotting and by real time RT-PCR after 10 -20 days of selection.
Western Blotting-To detect the expression of Myc-tagged NIC-1 and NIC-1 mutants, whole cell lysates were prepared in Nonidet P-40 lysis buffer (50 mM Hepes, pH 7.4, 1.5 mM EDTA, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40, 2 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, and 20 g/ml leupeptin). The lysates were cleared by centrifugation at 13,000 ϫ g for 30 min at 4°C. The supernatants were split into two aliquots and immunoprecipitated with either preimmune sera or anti-NIC-1 polyclonal antisera. Immune complexes were adsorbed to protein A-Sepharose, and eluted proteins were resolved by SDS-polyacrylamide gel electrophoresis on an 8% acrylamide gel. The proteins were transferred to an Immobilon P membrane (Millipore) and detected by immunoblotting with the anti-Myc tag polyclonal antibody (Upstate Biotechnology, Inc.; catalog number 06-549). CBF1 was detected by immunoblotting with anti-CBF1 polyclonal antisera raised against purified full-length CBF1.
Quantitative RT-PCR Analysis-Total RNA from K562 cells stably expressing the blank vector, wild-type NIC-1, or NIC-1 mutants was extracted with Trizol (Invitrogen). cDNA was synthesized by annealing RNA (2 g) with 250 ng of a 1:4 mixture of random and oligo(dT) primers by heating at 68°C for 10 min. After denaturation, the samples were incubated with Moloney murine leukemia virus reverse transcriptase (10 units/l; Invitrogen) combined with 20 mM dithiothreitol, 1 mM dNTPs, and 2 units/l RNasin (Promega) at 42°C for 1 h. The reaction mixture was heat-inactivated at 95-100°C for 5 min and diluted to a final volume of 200 l. Quantitative real time RT-PCR reactions (25 l) contained 2 l of cDNA, 12.5 l of SYBR Green (Applied Biosystems), and the appropriate primers. Product accumulation was monitored by SYBR Green fluorescence. The relative expression levels were determined from a standard curve of serial dilutions of cDNA samples. Forward and reverse primers for real time RT-PCR (5Ј-3Ј) were: human NIC-Myc, CCAGATCGCGCGCATC and GATATC-AGCTTCTGCTCCTCAGC; human hypoxanthine phosphoribosyl transferase, ATTGGTGGAGATGATCTCTCAACTTT and GCCAGTGCAAT-TATATCTTCCACAA; human Itch, AAACAGTCCCAAGTGGAAGCA and TCTGGTGACTCCACACACGAA; human Wwp1, CCAAAACCAC-TCGCATCTGAG and CGTGACAGACGCATTATCAGTTG; and human MMP1, ACAAATCCCTTCTACCCGGAA and CCCTTTGAA-AAAC-CGGACTTC.

Structure/Function Analysis of NIC-1 RAM Domain Internal
Deletion Mutants-Sequences within the highly conserved RAM domain (amino acids 1759 -1819) of NIC-1 and amino acids 2105-2114 are crucial for CSL-mediated transactivation and AP-1 repression (25). To test whether distinct sequences mediate activation versus repression and whether CSL binding is required for repression, we systematically mutagenized the RAM domain and conducted a structure/function analysis. Seven RAM domain internal deletion mutants were generated, which together with NIC-1(⌬1841-1847), resulted in mutagenesis of the majority of RAM domain sequences (Fig. 1). Expres-sion vectors encoding the mutants were transiently transfected into K562 erythroleukemia cells, and expression was assessed by immunoprecipitation of extracts with anti-NIC-1 antibody, followed by Western blotting with anti-Myc antibody. All of the mutants were expressed, and the expression levels did not differ greatly compared with wild-type NIC-1. However, NIC-1(⌬RAM) and NIC-1(⌬1759 -1778) were reproducibly expressed at higher levels than wild-type NIC-1 and other deletion mutants ( Fig. 2A).
To determine whether the mutants were competent to stably associate with endogenous CBF1, the blot of Fig. 2A was reprobed with anti-CBF1 antisera (Fig. 2B). Although CBF1 binding requires the RAM domain, the specific molecular determinants of binding have not been defined in detail. Endogenous CBF1 coimmunoprecipitated with wild-type NIC-1 and NIC-1(⌬1841-1847) but not the RAM domain deletion mutant NIC-1(⌬RAM). CBF1 coimmunoprecipitated with all RAM domain mutants except NIC-1(⌬1759 -1778), which lacks the first 20 amino acids of the N terminus of the RAM domain. Similar to NIC-1(⌬RAM), no CBF1 was recovered upon immunoprecipitation of NIC-1(⌬1759 -1778), whereas similar amounts of CBF1 were recovered upon immunoprecipitation of other RAM domain mutants and wild-type NIC-1.
The mutants were compared with wild-type NIC-1 for their capacity to activate transcription from a luciferase reporter plasmid containing four CBF1-binding sites (11). K562 cells were transiently transfected with the CBF1 reporter and a blank vector or expression vectors encoding wild-type or mutant NIC-1. Expression of wild-type NIC-1 strongly activated transcription of the CBF1 reporter, which binds endogenous CBF1 (Fig. 2C). Only the mutants that associate with CBF1 (Fig. 2B) activated the CSL luciferase reporter similar to wildtype NIC-1. NIC-1(⌬RAM) and NIC-1(⌬1759 -1778), which do not stably associate with CBF1, were almost completely defective in activating the reporter. The sequence in the middle shows the RAM domain, and the brackets denote sequences that were deleted in the constructs used in Fig.  2. The sequence at the bottom shows a subregion of the RAM domain, with the brackets denoting residues that were substituted with Ala residues to generate the constructs used in Fig. 3.
The NIC-1 mutants were also tested for their ability to repress AP-1-mediated transactivation. Endogenous AP-1 was activated by treating cells with the phorbol ester TPA. TPA strongly induced the activity of a reporter containing the AP-1-responsive collagenase 1 (MMP1) promoter with a single AP-1 site. Expression of the RAM domain mutants of NIC-1 that retained CSL binding capacity repressed AP-1 reporter activity similar to wild-type NIC-1 (Fig. 2D). NIC-1(⌬1759 -1778) and NIC-1(⌬RAM), which only weakly activated CSLmediated transactivation, only weakly repressed AP-1 (Fig. 2,  C and D). These results indicate that the first 20 amino acids of the RAM domain (1759 -1778) are essential for CBF1 binding and CSL-mediated activation and repression.
Structure/Function Analysis of NIC-1 RAM Domain Alanine Substitution Mutants-To pinpoint sequences mediating AP-1 repression, five additional mutants were generated, each with four alanine substitutions within the 20 amino acids of the RAM domain (Fig. 1). The expression of these mutants was assessed by immunoprecipitation with anti-NIC-1 antisera using extracts from transient transfected K562 cells. Western blot analysis with anti-Myc antibody revealed that all mutants except NIC-1(⌬RAM) and NIC-1(⌬1759 -1778) were expressed at levels similar to those of wild-type NIC-1 (Fig. 3A). The higher expression levels of NIC-1(⌬RAM) and NIC-1(⌬1759 -1778) suggest that amino acids 1759 -1778 may contain determinants of protein stability.

FIG. 2. CBF1 binding, CSL-mediated transactivation, and AP-1 repressing activities of NIC-1 RAM domain internal deletion mutants. A, Western blot analysis of wild-type NIC-1 and NIC-1 deletion mutants. A blank vector or a NIC-1 expression vector was introduced into K562 cells by transient
Plotting the level of activation and repression obtained from each mutant yielded a linear relationship with R 2 ϭ 0.816 (Fig.  4). Thus, CSL binding correlates with NIC-1-mediated repression of AP-1. Because the only known consequence of NIC-1 binding to CBF1 is to activate transcription of Notch target genes, these data are consistent with a model in which AP-1 repression requires activation of one or more target genes. The results with one mutant, NIC-1(⌬1841-1847), were not entirely consistent with this simple linear relationship. As noted in our previous study (25), NIC-1(⌬1841-1847) was partially compromised for activation but retained normal CBF1 binding and AP-1 repression. Thus, analogous to the 13 other mutants, the activity to repress AP-1 correlated with CBF1 binding. However, the efficiency of activation was modestly reduced upon deletion of amino acids 1841-1847. Linear regression analysis of activation versus repression data for all mutants except NIC-1(⌬1841-1847) yielded an R 2 value of 0.905.
A Powerful Cell-sensing Mechanism That Suppresses Protein Levels of NIC-1 and Transcriptionally Competent NIC-1 Mutants-Very low levels of NIC, below the threshold of detection by immunofluorescence, confer maximal CSL-mediated transactivation from reporter constructs in murine 3T3 cells (46). Based on this important finding, it is unclear whether only very low levels of NIC are generated upon Notch activation, whether cellular mechanisms suppress NIC levels, or whether cells repress transcription of ectopically expressed NIC. Wild-type NIC-1 and NIC-1 mutants are readily expressed upon transient transfection in K562 cells, and our previous work showed that stably transfected NIC-1 represses endogenous AP-1 target genes (25) and suppresses erythroid maturation of K562 cells (45). Studies were conducted to establish conditions to analyze the activity of the mutants studied in Figs. 2 and 3 in a chromosomal context in stably transfected K562 cells. Vectors encoding wild-type NIC-1, NIC-1 mutants, or an empty vector were transfected into K562 cells, and stably transfected pools of cells were selected with puromycin. Expression of wild-type NIC-1 and NIC-1 mutants was assessed by immunoprecipitation with anti-NIC-1 antisera using extracts from pools of stably transfected cells. Immunoprecipitated proteins were detected by Western blotting with anti-Myc antibody. Even though wild-type NIC-1 and NIC-1 mutants were expressed at comparable levels when transiently transfected into K562 cells ( Figs. 2A and 3A), wild-type NIC-1, NIC-1(1775-1778A), and NIC-1(⌬1841-1847), which retain transactivation/repression activities, were almost undetectable in the stably transfected cells. By contrast, mutants defective in both CSL-mediated transactivation and AP-1 repression (NIC-1(⌬RAM), NIC-1(⌬1759 -1778), NIC-1(1767-1770A), NIC-1(1771-1774A), and NIC-1(⌬2105-2114)) were highly expressed (Fig. 5, top (short exposure) and middle (long exposure) panels). ␣-Tubulin was expressed at similar levels in the extracts (Fig. 5, bottom  panel), indicating that the differential expression levels of wildtype NIC-1 and NIC-1 mutants are not accompanied by overt changes in protein expression. Thus, the expression analysis yielded the surprising finding that transcriptionally compro- mised NIC-1 mutants accumulate at high levels, whereas expression of transcriptionally active NIC-1 mutants is suppressed. Because both classes of NIC-1 mutants are expressed similarly in transiently transfected K562 cells, this suggests that a time-dependent cell-sensing mechanism suppresses the levels of NIC-1 and transcriptionally competent NIC-1 mutants.
The results of Fig. 5 can be explained by one of the following two mechanisms. Transcriptionally competent NIC-1 levels cannot be sustained, and therefore NIC-1 is degraded. Alternatively, expression vectors encoding transcriptionally competent NIC-1 are selectively repressed upon integration into chromosomal DNA. To distinguish between these mechanisms, total RNA was isolated from the same stably transfected K562 cells used for the protein analysis of Fig. 5. The expression of exogenous NIC-1 mRNA transcripts was quantitated by real time PCR (Fig. 6). Primers were used that amplify the C terminus of NIC-1 and the Myc tag to distinguish exogenous from endogenous NIC-1. The assays were conducted under linear conditions (Fig. 6A). Representative amplification curves for the detection of NIC-Myc and HPRT in K562 cells stably transfected with the empty vector (K562-Babe) or the NIC-1 expression vector (K562-NIC-1) are shown in Fig. 6B. Thermal dissociation of the PCR products revealed homogenous dissociation curves (Fig. 6C), indicative of a single major product for each primer set. Under conditions in which transcriptionally competent NIC-1 proteins were nearly undetectable (Fig. 5), the levels of RNA transcripts encoding NIC-1, transcriptionally competent NIC-1 mutants, and transcriptionally defective NIC-1 mutants were similar, differing by less than 2-fold. Thus, the very low protein expression levels of transcriptionally competent NIC-1 do not result from repression of the stably integrated expression vectors. Rather, a translational or post-translational mechanism suppresses the levels of active NIC-1 proteins in stably transfected K562 cells.
What mechanisms function translationally or post-translationally to strongly suppress the levels of transcriptionally competent NIC-1? NIC-1-dependent transcription might be tightly coupled to NIC-1 degradation, because this type of mechanism has been described for other transcriptional acti-vators (47)(48)(49). Sequences conferring degradation via the ubiquitin-proteosome system have been localized within transcriptional activation domains (48). Transactivation domains can signal to the ubiquitination machinery, and intriguingly, activator ubiquitination appears to be required for transactivation in certain contexts (50). Alternatively, transcriptionally competent NIC-1 might activate a gene(s) that directly mediates NIC-1 degradation.
Because wild-type NIC-1 and NIC-1 mutants were expressed at similar levels in transiently transfected cells (Figs. 2 and 3), the latter mechanism seems to be the most likely, unless the coupling of activation and proteolysis requires a chromosomal template. SEL-10, an F-box protein of the cdc4 family, negatively regulates Notch signaling and catalyzes ubiquitination and degradation of NIC-1 (51,52). However, SEL-10 interacts with the C-terminal portion of NIC-1, and the PEST domain present within this region is required for SEL-10mediated degradation of NIC-1 (53). Because mutations that greatly modulate NIC-1 protein levels (Fig. 5) do not occur within the C terminus of NIC-1 or within the PEST domain ( Fig. 1), it is unlikely that SEL-10 mediates the suppression of transcriptionally competent NIC-1 in stably transfected K562 cells.
Besides SEL-10, WW domain containing HECT domain ubiquitin ligases have been implicated in NIC-1 degradation. Drosophila Suppressor of Deltex down-regulates Notch signaling (54,55), whereas murine Itch binds and ubiquitinates NIC-1 in vitro (56). Although the molecular determinants of Itch binding to NIC-1 are unknown, Itch binds a NIC-1 derivative containing the RAM domain and the ankyrin repeats but lacking the PEST domain.
To test whether NIC-1 induces expression of Itch and/or the highly homologous ubiquitin ligase Wwp1, which act in a negative feedback loop to down-regulate NIC-1, we quantitated the levels of Itch and Wwp1 mRNA transcripts in K562 cells stably expressing NIC-1 by real time RT-PCR. This analysis revealed that expression of endogenous Itch and Wwp1 mRNA was unaffected by NIC-1 (Fig. 7). However, because NIC-1 strongly repressed TPA-mediated induction of MMP1 expression (Fig. 7), NIC-1 was functional in the cells.
These results indicate that the cell-sensing mechanism that suppresses the levels of transcriptionally competent NIC-1 in stably transfected K562 cells does not involve up-regulation of Itch or Wwp1 transcription. Our discovery of the differential expression of transcriptionally competent versus inactive NIC-1 derivatives in transiently versus stably transfected cells will provide a unique system for further elucidating mechanisms controlling Notch signaling via the regulation of NIC-1 protein levels.
In conclusion, we conducted a nonbiased, systematic analysis of the functionality of RAM domain sequences. Because the RAM domain of Notch shares no apparent sequence homology to known proteins, despite being highly conserved among Notch proteins, and secondary structure predictions fail to reveal structural motifs, logical predictions cannot be made regarding RAM domain structure/function. Surprisingly, the majority of RAM domain sequences were not required for CBF1 binding, CSL-mediated transactivation, and AP-1 repression (Figs. 2 and 3). Only eight amino acids within the N terminus of the RAM domain were essential for CBF1 binding, CSLmediated transactivation, and AP-1 repression (Fig. 8). Despite the high sequence conservation of amino acid residues at the extreme N terminus of the RAM domain (Fig. 8), these residues had no apparent functional role. The correlation between CBF1 binding and AP-1 repression and the inability of the mutants to differentially affect CSL-mediated transactivation and AP-1 repression strongly suggest that CBF1 binding mediates AP-1 repression. Because the only activity ascribed to the NIC-1-CBF1 complex is to regulate Notch target genes, the data are consistent with a model in which AP-1 repression requires activation of one or more Notch target genes. However, mechanisms in which CBF1 binding is required but induction of Notch target genes is not cannot be unequivocally ruled out, even though there is no precedent for this type of mechanism. The studies also led to the unexpected finding that NIC-1 activity is tightly coupled to NIC-1 expression levels. Although the experiments involved modulation of NIC-1 activity via mutations, cellular mechanisms that up-regulate or down-regulate NIC-1 activity might be similarly coupled to the control of NIC-1 protein levels. Such a relationship has major implications for the control of Notch signaling and therefore diverse developmental processes.
FIG. 8. A highly conserved subregion of the RAM domain containing amino acids essential for CBF1 binding, CSL-mediated transactivation, and AP-1 repression. Notch subtype sequences from various species were aligned. Gray shading denotes sequences identical to human Notch-1. The bold letters depict the amino acids that were shown in Figs. 2, 3, and 5 to be essential for CBF1 binding, CSL-mediated transactivation, AP-1 repression, and conferring very low level protein expression.