|Home | About | Journals | Submit | Contact Us | Français|
Identifying developmental processes regulated by Notch1 can be addressed in part by characterizing mice with graded levels of Notch1 signaling strength. Here we examine development in embryos expressing various combinations of Notch1 mutant alleles. Mice homozygous for the hypomorphic Notch112f allele, which removes the single O-fucose glycan in epidermal growth factor-like repeat 12 (EGF12) of the Notch1 ligand binding domain (lbd), exhibit reduced growth after weaning and defective T cell development. Mice homozygous for the inactive Notch1lbd allele express Notch1 missing an ~20 kDa internal segment including the canonical Notch1 ligand binding domain, and die at embryonic day ~E9.5. The embryonic and vascular phenotypes of compound heterozygous Notch112f/lbd embryos were compared with Notch1+/12f, Notch112f/12f, and Notch1lbd/lbd embryos. Embryonic stem (ES) cells derived from these embryos were also examined in Notch signaling assays. While Notch1 signaling was stronger in Notch112f/lbd compound heterozygotes compared to Notch1lbd/lbd embryos and ES cells, Notch1 signaling was even stronger in embryos carrying Notch112f and a null Notch1 allele.
Mouse embryos expressing the hypomorphic Notch112f allele, in combination with the inactive Notch1lbd allele which lacks the Notch1 ligand binding domain, died at ~E11.5-12.5. Notch112f/lbd ES cells signaled less well than Notch112f/12f ES cells but more strongly than Notch1lbd/lbd ES cells. However, vascular defects in Notch112f/lbd yolk sac were severe and similar to Notch1lbd/lbd yolk sac. By contrast, vascular disorganization was milder in Notch112f/lbd compared to Notch1lbd/lbd embryos. The expression of Notch1 target genes was low in Notch112f/lbd yolk sac and embryo head, whereas Vegf and Vegfr2 transcripts were increased. The severity of the compound heterozygous Notch112f/lbd yolk sac phenotype suggested that the allelic products may functionally interact. By contrast, compound heterozygotes with Notch112f in combination with a Notch1 null allele (Notch1tm1Con) were capable of surviving to birth.
Notch1 signaling in Notch112f/lbd compound heterozygous embryos is more defective than in compound heterozygotes expressing a hypomorphic Notch112f allele and a Notch1 null allele. The data suggest that the gene products Notch1lbd and Notch112f interact to reduce the activity of Notch112f.
Notch transmembrane receptors are important regulators of cell fate determination in numerous cell types [1-3]. Notch receptors in Drosophila and mammals are covalently modified with O-fucose on many epidermal growth factor-like (EGF) repeats of the extracellular domain . An important O-fucose site resides in epidermal growth factor-like repeat 12 (EGF12) which, together with EGF11, is required for canonical Notch ligand binding to Drosophila Notch [5-7] and to mammalian Notch1 . A point mutation that precludes the addition of fucose to EGF12 in Drosophila Notch results in enhanced binding of both Delta and Serrate Notch ligands, and a hyperactive Notch that is refractory to Fringe . However, the same mutation (Notch112f) in cultured mammalian cells gives markedly reduced signaling in a Notch reporter signaling assay [10,11], predicting a Notch1 null phenotype in vivo. Surprisingly however, homozygous Notch112f/12f mice are viable and fertile, but exhibit retarded growth and mild defects in T cell development in the thymus , consistent with weak Notch1 signaling. Notch1+/12f heterozygotes are indistinguishable from wild type in terms of growth and T cell development. However, compound heterozygotes carrying Notch112f and the inactive Notch1lbd allele, which lacks the ligand binding domain and generates an inactive ~280 kDa Notch1 receptor at the cell surface, are not born . Therefore Notch112f is a hypomorphic allele in mammals and the O-fucose glycan in the ligand binding domain is required for optimal Notch1 signaling. Homozygous Notch1lbd/lbd embryos die at ~E9.5 [8,12] with an indistinguishable phenotype compared to Notch1 null embryos (Notch1in32/in32 and Notch1tm1Con/tm1Con) described by others [13,14]. Heterozygous Notch1+/lbd and Notch1+/tm1Con mice are viable and fertile whereas Notch112f/lbd compound heterozygotes die between E11.5 and E12.5, significantly later than either Notch1lbd/lbd  or Notch1 null embryos [13,14] that do not express Notch1 [15-17].
The availability of these Notch1 mutant alleles suggested a genetic approach to determining effects of varying Notch1 signaling strength. The Notch1lbd mutation generates a non-functional but cell surface-expressed Notch1 that cannot signal [8,12]. Notch1tm1Con lacks Notch1 on the cell surface due to the absence of its transmembrane domain . Notch1in32 homozygous embryos have no Notch1 transcripts  and an indistinguishable phenotype from Notch1tm1Con homozygotes which lack Notch1 based on western analyses [15,18]. Notch1+/- heterozygotes carrying either of the Notch1 null alleles exhibit Notch1 signaling defects in certain cell types, an effect attributed to Notch1 haploinsufficiency rather than to a dominant negative effect in Notch1tm1Con [18-21]. In this paper we compare embryogenesis and vasculogenesis in compound heterozygotes expressing the hypomorphic Notch112f allele with either the inactive Notch1lbd allele [8,12] or the Notch1tm1Con null allele .
The Notch112f and Notchlbd alleles investigated here are diagrammed in Fig. Fig.1A1A and and1B1B and their identification by PCR genotyping is shown in Fig. Fig.1C.1C. Previous studies showed that Notch112f/lbd compound heterozygotes die by ~E12.5 . To examine Notch ligand binding and the strength of Notch signaling in more detail, ES cells were derived from Notch112f/lbd compound heterozygous blastocysts and compared to ES cells derived from Notch112f/12f and Notch1lbd/lbd homozygous blastocysts and wild type ES cells (Fig. (Fig.2).2). All cell lines bound the anti-Notch1 extracellular domain mAb 8G10 equivalently, and therefore expressed the various Notch1 molecules similarly at the cell surface (Fig. (Fig.2A).2A). Each mutant line exhibited a decrease in the low level of soluble Delta1 binding observed with wild type ES cells (Fig. (Fig.2B).2B). Binding of Delta1 is not reduced to zero even in Notch1 null ES cells because of the presence of Notch2, Notch3 and Notch4 . Notch signaling was analysed in co-culture assays with L cells or L cells expressing full length Delta1 or Jagged1 ligand. This reporter assay revealed a graded reduction in Notch signaling with Notch112f/12f >Notch112f/lbd >Notch1lbd/lbd ES cells (Fig. 2C-D). This graded response was also observed by western analysis using Notch1 antibody Val1744  which detects the ~110 kDa Notch1 fragment generated by γ-secretase complex cleavage of Notch1. The level of activated Notch1 in Notch112f/lbd ES cells was less than in Notch112f/12f ES cells, which was lower than in control ES cells, while Notch1lbd/lbd ES cells had undetectable levels of activated Notch1 (Fig. (Fig.2E).2E). Nevertheless, all ES cell populations, including Notch1lbd/lbd ES cells, expressed equivalent levels of full-length Notch1 (Fig. (Fig.2E).2E). Taken together, these data indicate that Notch112f and Notch1lbd expression and transit to the cell surface were similar to wild type Notch1, but Notch1 signaling was reduced in mutant cells: Notch112f signaling was sightly less than wild type; signaling from the combination of Notch112f and Notch1lbd was further reduced, and signaling by Notch1lbd alone was essentially absent. Previous experiments have shown that Notch1lbd/lbd and Notch1in32/in32 ES cells which lack Notch1 [13,15,16], are equally deficient in Delta1-Fc binding and Notch1 signaling .
Embryonic development was compared between Notch112f/12f, Notch112f/lbd and Notch1lbd/lbd embryos. At E9.5 Notch112f/lbd embryos formed 17-21 somites compared to 23-26 somites in Notch112f/12f embryos, the same as Notch1+/+ embryos, and 13-17 somites in Notch1lbd/lbd embryos , the same as Notch1tm1Con null embryos  (Table (Table1).1). Compared to Notch112f/12f and Notch1+/12f embryos, Notch112f/lbd embryos also showed severely defective vasculogenesis in yolk sac at E9.5, similar to Notch1lbd/lbd yolk sac. By contrast, Notch112f/lbd embryos at E9.5 and E10.5 exhibited milder defects in development than Notch1lbd/lbd embryos  (Fig. (Fig.3),3), although the ballooning of the pericardial sac and defective heart development were severe, and similar to mutants globally defective in Notch signaling such as mutants lacking Pofut1 , RBPJk  or presenilins 1 and 2 . Taken together, these data indicate that two copies of Notch112f do not noticeably affect mouse embryogenesis at a gross level, whereas a single copy of Notch112f with Notch1lbd support embryonic development ~2.0-2.5 days longer than embryos with two copies of Notch1lbd.
Notch1 signaling is critical for vasculogenesis during mouse embryogenesis . Loss of Notch1 in embryos  or in endothelial cells  causes embryonic lethality with severe vascularization defects in yolk sac, placenta and embryo. Blood that had leaked from the heart and blood vessels was apparent in Notch112f/lbd embryos (Fig. 3I-K; arrows). Vascular organization in embryos was examined by staining with anti-Pecam1 (endothelial marker platelet/endothelial cell adhesion molecule-1). Notch112f/12f embryos (Fig. 4B, F, J, N) did not exhibit any apparent defects in brain, heart or intersomitic vascularization compared to Notch1+/12f embryos. Notch112f/lbd embryos exhibited somewhat disorganized vascularization in embryos, especially in the main trunk of the anterior cardinal vein, the vascular network of the head and heart, and in intersomitic vessels (Fig. 4C, G, K, O). Notch1lbd/lbd embryos exhibited severe defects in vascularization (Fig. 4D, H, L, P). Therefore, the extensive vascularization in E9.5 and older Notch112f/lbd embryos appears to be well supported by the level of Notch1 signaling provided by the Notch112f allele. Considering that the vascular defects in yolk sac of compound heterozygous Notch112f/lbd and homozygous Notch1lbd/lbd embryos were similarly severe, the comparatively milder defects in Notch112f/lbd embryos indicated that Notch1 signaling from a single copy of Notch112f, while not sufficient to support vascularization in yolk sac at E9.5, is able to support a high level of vascularization in E9.5 embryos. It seems that vascularization in yolk sac requires stronger Notch1 signaling than in the embryo.
Whereas vascularization was severly affected in both yolk sac and embryo of Notch1lbd/lbd embryos, only the yolk sac of Notch112f/lbd compound heterozygous embryos exhibited extremely defective vascularization. To investigate further, the expression of vasculogensis-related and Notch1 target genes was examined by real-time PCR using total RNA isolated from E10.5 Notch112f/lbd and Notch1+/12f yolk sacs and embryo heads. The relative expression levels of Pecam1 and Vegf were increased in Notch112f/lbd yolk sacs and embryos, and Vegfr2 expression was increased in Notch112f/lbd embryo heads (Fig. 5A-C). Therefore loss of Notch1 signaling upregulated transcription of the Pecam1, Vegf and Vegfr2 genes. Interestingly, the increased expression of Vegf and Vegfr2 was greater in Notch112f/lbd embryos, consistent with the relative strength of Notch1 signaling being greater in yolk sac. Expression of the Notch1 target genes Hes5, Hey1 and Hey2 was reduced in Notch112f/lbd yolk sac (Fig. 5D-F), but the level of Hes1 transcripts was not changed (data not shown). In embryos, only the expression of Hes5 was significantly reduced compared to control. The expression of Ang1, Tie2 and Ephrin-B2 which are involved in angiogenesis, as well the expression of Notch1 itself, were not changed in Notch112f/lbd yolk sac or embryos (data not shown). The fact that the increase in Vegf and Vegfr2 transcripts was more in embryo head than yolk sac (418% vs 170% for Vegf; 227% vs. 148% for Vegfr2; Fig. Fig.5B5B and and6C),6C), and the fact that the reduction in Notch target gene expression was greater in yolk sac than embryo head, correlated generally with Notch1 signal strength and the greater severity of vascularization defects in yolk sac versus embryo head.
The severity of the Notch112f/lbd phenotype suggested an interaction between Notch112f and Notch1lbd that interfered with signaling by Notch112f. In this case, compound heterozygous embryos expressing a Notch112f allele and a Notch1 null allele might be expected to have a milder phenotype than Notch112f/lbd embryos. Notch112f/12f mice were crossed with Notch1+/tm1Con heterozygotes and embryos were examined at E9.5 and later (Fig. (Fig.6,6, Table Table2).2). Some Notch112f/tm1Con embryos died between E11.5 and E12.5 with similar defects to Notch112f/lbd embryos. However, this is ~1.5 days later than observed with Notch1tm1Con/tm1Con homozygous embryos who were mostly dead by E10 . However, nearly one third of the Notch112f/tm1Con embryos developed beyond E12.5 and died at various times during embryogenesis, including after birth (Table (Table2).2). Two Notch112f/tm1Con pups were found after birth, but none were observed after postnatal day 7 (Fig. (Fig.6,6, Table Table2).2). Somite numbers in Notch112f/tm1Con embryos varied from as low as Notch112f/lbd embryos to as high as wild type embryos (Table (Table1).1). Taken together, these results indicate that Notch112f receptors present at a 50% dose in vivo, generate stronger Notch1 signaling than Notch112f in combination with Notch1lbd. This provides genetic evidence that Notch112f and Notch1lbd may functionally interact.
In this paper we show that Notch1 signaling is greatly reduced in Notch112f/lbd ES cells and compound heterozygous embryos, but is significantly greater than in Notch1lbd/lbd ES cells or homozygous embryos. The presence of the hypomorphic Notch112f allele allows vasculogenesis to proceed further and embryos to survive ~1.5-2 days longer. The vascular system develops early during mammalian embryogenesis. Initially, endothelial cell precursors differentiate and coalesce into a primitive network of undifferentiated blood vessels (the primary vascular plexus) in both the mammalian embryo and its extraembryonic membrane the yolk sac, in a process termed vasculogenesis . Subsequently, the primary vascular plexus is remodeled into a highly organized and functionally competent vascular network in a process termed angiogenesis [29,30]. These processes are controlled by several signaling molecules, including vascular endothelial growth factor (VEGF) and its receptors , angiopoeitin 1 and its receptor Tie2 , Ephrin-B ligands and EphB receptors , TGFβ and its receptors , and Notch receptors and their ligands Delta and Jagged [25,26,35-38]. Defects in vasculogenesis are one of the major reasons that Notch1 null embryos die at mid-gestation [13,26]. Conditional mutation of Notch1 in vascular endothelial cells using the Tie2-Cre transgene showed that embryos lacking endothelial cell Notch1 die at ~E10.5 with profound vascular defects in placenta, yolk sac, and the embryo proper . The Notch1 target genes Hey1 and Hey2 are also essential for embryonic vascular development . A requirement for Notch signaling in the maintenance of vascular homestasis and the repression of endothelial cell proliferation is also indicated in adult mice by conditional deletion of RBP-Jκ in endothelial cells .
Interestingly, Notch112f/lbd embryos allowed us to observe that vasculogenesis is regulated to different extents in yolk sac and embryo by Notch1 signaling. Thus, vascular defects in Notch112f/lbd yolk sac were as severe as in Notch1lbd/lbd yolk sac, but vascular defects in Notch112f/lbd embryo heads were comparatively mild. The vasculogenic phenotype of Notch112f/lbd embryos was also milder than reported for Jagged1 or Notch1 or Notch1/4 null embryos [13,26,37], reflecting the presence of a low level of Notch1 signaling in Notch112f/lbd compound heterozygotes. The reduced strength of Notch1 signaling was responsible for defective artery development in Delta-like 4 (Dll4) heterozygous embryos . Hes5 and Hey1 are Notch1 target genes, and Notch1 downregulates expression of Hesr-1/Hey1 thereby enhancing expression of its target gene Vegfr2 in endothial cells . In addition, Vegf is upregulated six-fold in Hey1/2 double knock-out embryos . Notch1 has also been proposed to regulate vasculogenesis and angiogenesis via induction of Ephrin-B2 [42,43] and Ang1 [44,45], and suppression of Vegfr-2/Kdr [41,46]. Consistent with this, we observed enhanced suppression of Vegfr2 and Vegf in Notch112f/lbd yolk sac and embryo. However, we observed no change in the expression of Ang1, Tie2, Ephrin-B2 or Notch1 itself, although experiments in human endothelial cells indicate that Ang1 and Tie2 are Notch1 target genes [44,45]. Ephrin-B2 was reported to respond to Notch4, but not to Notch1, through Delta-like 4 in differentiating HUVEC cells , so it was perhaps not surprising that Ephrin-B2 expression was unchanged in Notch112f/lbd yolk sac or embryo. Thus, decreased Notch1 signaling may inhibit vascular development in yolk sac more than in embryos by inducing more Vegf and Vegfr2 through generating less Hes5 and Hey1 mRNA in yolk sac.
The prolonged embryonic development supported by the hypomorphic Notch112f allele was only ~1.5-2 days for Notch112f/lbd embryos compared to Notch1lbd/lbd [8,12], or Notch1 null embryos [13,14]. By contrast Notch112f/12f, Notch1+/lbd or Notch1+/- heterozygotes are viable and fertile [12-14,20]. This suggests that Notch1lbd may interfere with Notch112f in a process termed negative complementation for Abruptex Notch mutants in Drosophila [47,48]. The basis of negative complementation is most commonly attributed to the products of the mutant alleles interacting physically . Thus Notch1lbd may either be dominant negative and inhibit Notch112f activity, or may not form a functional dimer or higher oligomer with Notch112f, if that is required for Notch1 to function. We prefer the latter hypothesis because there is no evidence to date that Notch1lbd behaves as a dominant negative in Notch1+/lbd heterozygotes [8,12]. Unfortunately, attempts to prove the existence of dimers or higher oligomers of Notch1 expressed at endogenous levels have so far been unsuccessful and previous attempts came to opposite conclusions. While two groups found that overexpressed Notch1 transfected into cultured cells may form dimers through the transmembrane domain or the extracellular domain EGF repeats, one group concluded that dimerization is necessary for Notch1 to signal , while the other concluded that Notch1 signals without the need for dimerization, and is present mainly as a monomer on the cell surface . Both studies characterized transiently-transfected Notch1 expressed at much higher levels than endogenous Notch1, which might induce anomolous interactions.
If Notch1lbd reduces the effective amount of Notch112f to a level insufficient to sustain development, we reasoned that Notch112f expressed in the context of a Notch1 null background may function better. In fact, we found that a significant proportion of Notch112f/tm1Con embryos survived beyond E12.5 and that some survived to birth. On the other hand, some compound heterozygous Notch112f/tm1Con embryos died at ~E11.5 with similar defects to Notch112f/lbd. This indicates that Notch112f at a dose of 50% functions at a threshold of Notch1 signaling strength that variably sustains embryogenesis through to birth - a stochastic effect or perhaps a genetic background effect, since Notch1+/12f and Notch1+/lbd mice were not extensively backcrossed to C57Bl/6. Nevertheless, the Notch1 signal strength generated by a single copy of Notch112f was intermediate between Notch112f/12f and Notch112f/lbd, revealing the importance of maintaining a certain level of Notch1 signaling for mouse embryogenesis to proceed. Fig. Fig.77 summarizes these findings in a diagram which describes a mini-allelic series of available Notch1 mutants. It includes the Notch1 processing point mutant Val1744Gly (Notch1v!g/v!g) which has a phenotype very similar to, but slightly less penetrant than, a Notch1 null . It also includes Notch1+/null heterozygotes that have mild Notch1 signaling defects uncovered in competition assays  or by close examination of specific cell types [18,20,21]. Haploinsufficiency of NOTCH1 is the basis of aortic valve disease in humans . We predict that Notch1+/lbd and Notch1+/12f heterozygotes have slightly less Notch1 signaling than Notch1+/tm1Con and should display evidence of more extensive Notch1 signaling defects in particular cell types. The range of Notch1 mutant alleles available in the mouse should be helpful in identifying new in vivo functions for Notch1.
Mice carrying Notch1 lacking the O-fucose site in EGF12 (Notch112f) and mice carrying Notch1 lacking the ligand binding domain (Notch1lbd) were generated by gene targeting as previously described [8,12]. They were backcrossed 6-7 generations to C57/Bl6 mice before being used in these experiments. Notch1l2f/lbd embryos were obtained by crossing Notch112f/12f and Notch1+/lbd mice. Embryos were collected from E9.5 and yolk sac DNA was genotyped by PCR using primers 5F: GTATGTATATGGGACTTGTAGGCAG and 6R: CTATGAGGGGTCACAGGACCAT, that give a 466 bp product for the Notch112f allele and a 363 bp product for the wild type Notch1 allele; and primers 5F and 9R: CTTCATAACCTGTGGACGGGAG that give a 575 bp product for the Notch1lbd allele. The Notch1 null allele (Notch1tm1Con) encoding Notch1 lacking the transmembrane domain  backcrossed extensively to C57Bl/6 was kindly provided by Cynthia Guidos, University of Toronto. Notch112f/tm1Con embryos were obtained by crossing Notch112f/12f and Notch1+/tm1Con mice and genotyped by PCR using primers neo-F: CTTGGGTGGAGAGGCTATTC and neo-R: AGGTGAGATGACAGGAGATC for the Notch1tm1Con allele and primers loxF: GGCGAGCTCGAATTGATCC and 9R for Notch112f allele. Mice were housed under conventional barrier protection in accordance with Einstein and NIH guidelines. Protocols were approved by the Albert Einstein Animal Institute Committee.
ES cells were isolated from E3.5 blastocysts as described , and genomic DNA was genotyped by PCR as described above. ES cells were routinely cultured on an SNL2 γ-irradiated feeder layer  in DMEM supplemented with 15% fetal bovine serum (Gemini, West Sacramento, CA), non-essential amino acids, L-glutamine, 1000 U ESGRO® (Chemicon, Temecula, CA), 1% β-mercaptoethanol, 25 mM HEPES, penicillin (50 U/ml) and streptomycin (50 μg/ml). All reagents were from SpecialtyMedia, Lavellette, NJ. Before use in experiments, ES cells were passaged on gelatinized plates for 2-3 generations to remove feeder cells.
ES cells cultured on gelatinized plates were lysed in RIPA buffer (Upstate, Lake Placid, NY) containing complete protease inhibitor cocktail (Roche, Basel, Switzerland) for 30 min on ice and debris was removed by low speed centrifugation. Lysates were resolved by SDS-PAGE, transferred to polyvinyldifluoride (PVDF) membrane and probed with 8G10 anti-Notch1 mAb (Upstate, 57-557, 1:500, Lake Placid, NY) for full-length Notch1 or Val1744 Notch1 antibody (Cell Signaling Technology, Val1744, 1:1000, Beverly, MA) for cleaved, activated Notch1, followed by horseradish peroxidase(HRP)-conjugated secondary antibodies. Reactive bands were visualized with Enhanced Chemiluminescence Reagent (Amersham Pharmacia Biotech, Piscataway, NJ). β-tubulin-III specific antibody T8660 (Sigma Chemical Co., St. Louis, MO) was used as a loading control.
For cell surface Notch1 expression, 70-80% confluent ES cells were dissociated from plates using phosphate-buffered saline (PBS)-based enzyme-free dissociation solution (SpecialtyMedia, Lavellette, NJ) for 10 min at 37°C. After washing, ES cells (5 × 105) were incubated with 0.5 μg 8G10 anti-Notch1 antibody in Hank's balanced salt solution containing 3% bovine serum albumin Fraction V (Sigma Chemical Co., St. Louis, MO), 1 mM CaCl2 and 0.05% Na azide (HBSS/BSA) for 1 h at 4°C, washed and incubated in Alexa-488 conjugated anti-Hamster IgG (1:100) in HBSS/BSA in the dark (Invitrogen, Carlsbad, CA) for 30 min at 4°C. Immunofluorescence was analyzed on a FACSCalibur flow cytometer (BD Biosciences, San Diego, CA), gating on live cells determined by 7-AAD staining. Data were analyzed using Flowjo software (Tree Star, San Carlos, CA).
Notch signaling assays were performed in duplicate as previously described [54,55]. ES cells were plated at 2 × 105 cells per well of a six-well plate in ES medium, and co-transfected the next day with 0.2 μg of TP1-luciferase Notch reporter plasmid and 0.05 μg of Renilla luciferase reporter (pRL-TK; Promega, Madison, WI) along with 1.8 μg empty vector alone using FuGene 6 (Roche, Basel, Switzerland). At 16 h post-transfection, ES cells were overlaid with 1 × 106 rat Jagged1-expressing L cells (Jagged1/L), Delta1-expressing L cells (Delta1/L) or parental L cells . At 48 h after transfection, firefly and Renilla luciferase activities were quantitated in cell lysates using a dual luciferase assay (Promega, Madison, WI). Ligand-dependent Notch activation was expressed as relative fold-activation of normalized luciferase activity stimulated by ligand/L cells compared to L cells.
Soluble Notch ligand Delta1 with human Fc tag [57,58] was prepared form HEK-293T cells expressing Delta1-Fc  cultured in α-MEM containing 10% FBS until 70~80% confluence. The medium was changed to 293 SFM II serum-free medium (Invitrogen) and conditioned medium was collected after 3 days. Cellular debris was removed by low-speed centrifugation, the supernatant was filtered and stored at 4°C. Soluble ligand concentration was determined by western blotting using HRP-conjugated anti-human IgG antibody (Jackson Immunoresearch, West Grove, PA). For the binding assay, ES cells on plates were dissociated using PBS-based Enzyme-free dissociation medium for 10 min at 37°C, and the single cell suspension of ES was incubated with 2 μg/ml Delta1-Fc in HBSS/BSA for 1 h at 4°C, followed by incubation with 1:100 phycoerythrin (PE)-conjugated anti-human Fc antibody (Jackson Immunoresearch, West Grove, PA) for 30 min at 4°C. After washing, live cells determined by gating on the 7-AAD negative population were analyzed on a FACS Calibur flow cytometer (BD Biosciences, San Jose, CA). Ligand binding ability was measured as mean fluorescence intensity (MFI) using Flowjo software (Tree Star, San Carlos, CA).
Embryos were collected on E9.5 and DNA from yolk sac was genotyped by PCR. Embryos were fixed in 4% paraformaldehyde (PFA) in PBS overnight at 4°C, dehydrated through a methanol series, and bleached in 5% H2O2/methanol for 5 h. Embryos were rehydrated and placed in PBSMT (PBS containing 3% nonfat milk, 0.1% Triton X-100). After 2 h, embryos were incubated with anti-mouse Pecam1 (1:200; BD Biosciences, San Jose, CA) in PBSMT overnight at 4°C. After 5 washes with PBSMT embryos were incubated in a 1:200 dilution of HRP-conjugated secondary antibody (Zymed, South San Francisco, CA) overnight. Embryos were washed 5 times in PBSMT and rinsed in PBT (PBS containing containing 0.2% BSA, 0.1% Triton X-100), followed developing with DAB kit (Vector Laboratories, Burlingame, CA). Finally, embryos were washed in PBT and postfixed in 4% PFA, dehydrated through a methanol series and cleared in BABB (benzyl alcohol: benzyl benzoate - 1:2) in a glass Petri dish. Photos were taken in PBS or BABB using an inverted phase contrast microscope (Olympus IMT-2, Olympus America Inc., Center Valley, PA) and a Canon S40 camera with T-mount adaptor.
Total RNA was extracted from yolk sac or embryo head using TRIZOL® reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Aliquots of 1 μg RNA were digested by DNase I and cDNA was prepared using RNA PCR Kit ver. 3.0 (Takara Mirus Bio, Madison, WI) with oligo dT priming. Real-time PCR reactions with SybrGreen quantification were established with 1/20 of each cDNA preparation in an Opticon2 DNA Engine (MJ Research, Cambridge, MA). Relative expression levels after normalization using β-actin were calculated using the 2-ΔΔCT method ()( and confirmed by the absolute quantification method using standard curves. Primer pairs for real-time PCR were Ang1 (CATTCTTCGCTGCCATTCTG, GCACATTGCCCATGTTGAATC), Pecam1 (GAGCCCAATCACGTTTCAGTTT, TCCTTCCTGCTTCTTGCTAGCT) , Vegf (GGAGATCCTTCGAGGAGCACTT, GCGATTTAGCAGCAGATATAAGAA), Tie2 (ATGTGGAAGTCGAGAGGCGAT, CGAATAGCCATCCACTATTGTCC), Hey1 (TGAGCTGAGAAGGCTGGTAC, ACCCCAAACTCCGATAGTCC), Hey2 (TGAGAAGACTAGTGCCAACAGC, TGGGCATCAAAGTAGCCTTTA), Ephrin-B2 (GCGGGATCCAGGAGATCCCCACTTGGACT, GTGCGCAACCTTCTCCTAAG), Hes1 (AAGGCGGACATTCTGGAAAT, GTCACCTCGTTCATGCACTC) . Hes5 (TACCTGAAACACAGCAAAGC, GCTGGAGTGGTAAGCAG)  and β-actin (GTGGGCCGCTCTAGGCACCA, TGGCCTTAGGGTTCAGGGGG). All real-time PCR experiments were performed in duplicate from ≥ 4 independent samples.
Statistical significance was calculated using the unpaired Student's t-test (two-tailed) using Graphpad Prism (GraphPad Software, Inc., San Diego, CA) unless otherwise noted.
PS conceived the project, obtained funding, participated in the design of experiments and analysis of data, and co-wrote the manuscript; CG partipated in the design of the experiments, performed or participated in all experiments, analysed data and co-wrote the paper. All authors read and approved the final version of the manuscript.
We thank Wen Dong for excellent technical assistance, Linchao Lu for helpful suggestions and Bin Zhou for helpful comments on the manuscript. This work was supported by NIH grant NCI RO1 95022 to PS and in part by NCI grant PO1 13330 to the Albert Einstein Cancer Center.