To generate mice with Notch1 lacking the putative ligand binding domain, embryonic stem (ES) cells with loxP sequences flanking exons 6 – 8 of mouse Notch1 were generated by gene targeting (Fig. 1A; 13). Exons 7 and 8 encode EGF11 and EGF12 and exon 6 was included in order that the mutant Notch1 was ~20 kDa lower in molecular weight. Two independent ES colonies selected for resistance to G418 were shown by Southern analysis to carry a targeted Notch1 allele (Fig. 1B). Chimeric mice carrying the mutant allele were crossed with mice expressing the MeuCre40 recombinase transgene 14 to obtain mice with a Notch1^lbd allele after deletion of exons 6 – 8 along with the HSVtk/Neo cassette (Fig. 1A). Southern blot and PCR analysis of genomic DNA were used to genotype E9.5 embryos of Notch1^+/lbd crosses (Fig. 1C). All expected genotypes were represented at this stage. However, only wild type and heterozygous pups were born from 6 litters (Table 1).

To determine when Notch1^lbd/lbd embryos die, embryos from Notch1^+/lbd crosses were examined during embryogenesis. Notch1^lbd/lbd embryos were indistinguishable from wild type at ~E8.75, but by ~E9.5 Notch1^lbd/lbd embryos were severely growth-retarded, with a tube-like heart, distended pericardial sac, and defective vascularization of the yolk sac (Fig. 1D). By ~E10.5, many mutant embryos were resorbed and all mutant embryos were resorbed by ~E11.5. Therefore Notch1^lbd/lbd embryos exhibited global defects in Notch signaling with a phenotype indistinguishable from Notch1ⁱⁿ³² 8 or Notch1^tm1/Con/1 9 null embryos.

Blastocysts from heterozygous Notch1^+/lbd crosses were used to isolate embryonic stem (ES) cell lines of each genotype. Reverse transcription (RT)-PCR of total RNA showed that Notch1^lbd/lbd ES cells expressed Notch1 transcripts at levels similar to wild type and heterozygous ES cells (Fig. 1E). Notch1^lbd/lbd ES cells had a similar growth rate to Notch1^+/lbd and Notch1^+/+ ES cells (Fig. 2A). This was also observed with Notch1 null ES cells 10. An antibody to the extracellular domain of Notch1 (8G10) detected the ~300 kD full length Notch1 in wild type ES cells and the ~280 kDa truncated Notch1 in Notch1^lbd/lbd ES cells (Fig. 2B). The ~180 kDa Notch1 extracellular domain was not routinely observed, but when present it was in similar amounts in Notch1^lbd/lbd and Notch1^+/+cells. Flow cytometry showed that equivalent amounts of wild type and mutant Notch1 receptors were present on the surface of Notch1^+/+ and Notch1^lbd/lbd ES cells, respectively (Fig. 2C). Therefore the internal deletion that includes the putative ligand binding domain did not affect Notch1 stability or trafficking to the cell surface.

When Notch1 binds canonical Notch ligands, it undergoes cleavage by γ-secretase and the new N-terminus of activated Notch1 may be detected by the antibody Val1744 15. Western blot analysis revealed a robust signal for activated Notch1 in cultured wild type ES cells but no corresponding signal was observed in Notch1^lbd/lbd ES cells (Fig. 2B). Thus while Notch1^lbd is expressed at the cell surface it is not activated under conditions that activate wild type Notch1, presumably because of the loss of its ligand binding domain. Indeed the Notch ligand Delta1 had reduced binding to Notch1^lbd/lbd ES cells (Fig. 2D). ES cells express each of the four mammalian Notch receptors and all would be expected to bind Delta1. To confirm that the reduced binding of Delta1 to Notch1^lbd/lbd is due to the Notch1^lbd mutation, we examined Delta1 binding to Notch1 null ES cells (Notch1^in32/in32) termed 290-2 which lack Notch1 on the cell surface 10. The binding of Delta1 to 290-2 cells was reduced to the same extent as to Notch1^lbd/lbd cells (Fig. 2D). Therefore, the deletion of EGF repeats 8–12 in mouse Notch1 eliminates Delta1 binding to Notch1 as expected, but binding to other Notch receptors remains. This residual binding was prevented by including EDTA in the binding buffer under conditions that prevent Notch/ligand binding but do not release Notch receptors from the cell surface. Delta1-induced Notch1 signaling was also reduced in Notch1^lbd/lbd ES cells as shown by a co-culture reporter assay which detects signaling through all four Notch receptors (Fig. 2E). A second canonical Notch ligand, Jagged1, was also defective at inducing Notch signaling in Notch1^lbd/lbd ES cells (Fig. 2E). Delta1- and Jagged1-induced Notch signaling was also reduced, but not eliminated, in Notch1^in32/in32 ES cells which are Notch1 null 10 (Fig. 2E). Residual Notch signaling presumably reflects the presence of the other three Notch receptors.

The Notch1^lbd mutant allele is transcribed similarly to the wild type allele in Notch1^lbd/lbd ES cells (Fig. 1E) but Notch1^lbd/lbd ES cells are defective in Notch1 signaling (Figs. 1 and 2). The Notch1^lbd mutant allele therefore allowed us to determine if signaling via Notch1 is required for the Notch1 gene to be expressed in vivo. Notch1^+/lbd females were crossed with Notch1^+/lbd males and embryos were examined at mid-gestation. Notch1^lbd/lbd embryos were morphologically indistinguishable from wild type embryos at E8.75 and were examined for Notch1 gene expression by in situ hybridization. It can be seen in Fig. 3A that Notch1 was expressed in the presomitic mesoderm (PSM) the forming somite (arrows) and the neural tube. The hybridization signal obtained for wild type and Notch1^lbd/lbd embryos was similar. Therefore expression of the Notch1 gene was not markedly altered by the loss of canonical Notch1 signaling in early embryogenesis. By contrast, following removal of Pofut1 16 or Mind bomb 1 1718, both of which inhibit Notch signaling through all four Notch receptors, Notch1 expression is reduced in the PSM and enhanced in neural tube and mesencephalon of E8.75 embryos. At E9.0, Notch1 transcripts were markedly reduced in Notch1^lbd/lbd embryos (Fig. 3B). By E9.5 when Notch1^lbd/lbd embryos were dying, Notch1 gene expression was severely reduced (Fig. 3C).

Whole mount in situ hybridization provided additional confirmation that the Notch1^lbd mutation inactivates Notch1 signaling. The Notch1 target and somitogenic genes Hes5, Myogenin and Uncx4.1 were examined in embryos at E9.5. At that stage Hes5 is expressed in neural tube, brain and the forming and formed somites. The Hes5 gene is a known target of Notch1 signaling and its expression was severely reduced in Notch1^lbd/lbd embryos (Fig. 3D). Myogenin is expressed in mature somites of wild type embryos at E9.5. In Notch1^lbd/lbd embryos which had 13–17 (n = 3) poorly-formed somites, Myogenin expression was greatly reduced (Fig. 3E). The myogenic transcription factor Uncx4.1 is expressed on the posterior side of mature somites and in the PSM of wild type embryos at E9.5. In Notch1^lbd/lbd embryos, expression in somites and PSM was lost (Fig. 3F). However, expression of Uncx4.1 was induced in brain in the absence of Notch1 signaling, as observed previously in embryos defective in signaling through all four Notch receptors 16. We previously showed that cyclin D1 expression is markedly reduced in Notch1^lbd/lbd embryos 19. Importantly therefore, the Notch1 signaling defects observed in Notch1^lbd/lbd embryos are not rescued by non-canonical Notch1 ligands that might bind to the large portion of the Notch1 extracellular domain that remains in Notch1^lbd.

Following oocyte-specific deletion of Pofut1 or RBP-Jκ null oocytes are fertilized and mutant blastocysts develop through gastrulation 2021. Pofut1 16 and RBP-Jκ 22 are essential for Notch signaling through all four Notch receptors. Pofut1 transfers fucose to Notch receptors and RBP-Jκ complexes with the cleaved ICD of all Notch receptors, but both activities might have effects that are independent of the Notch pathway. Thus it was of interest to determine if Notch1^lbd/lbd blastocysts could develop and implant because Notch1 is expressed in oocytes, fertilized eggs and blastocysts 1112. To obtain maternal and zygotic mutant blastocysts, females homozygous for the Notch1 floxed allele (Fig. 1A) and carrying a ZP3Cre transgene were generated (Fig. 4A). Notch1^F/F:ZP3Cre females were mated with Notch1^+/lbd or wild type males. Pups and E9.5 embryos were genotyped (Table 2). At birth, all pups from Notch1^F/F:ZP3Cre by wild type crosses were heterozygous showing that the Cre recombinase was highly efficient since the Notch1^F allele was not transmitted. At E9.5, 29 embryos from 4 crosses included 15 mutants (Notch1^lbd/lbd) and 14 heterozygotes (Notch1^+/lbd). Therefore, eggs with Notch1 lacking the ligand binding domain were fertilized by sperm that also lacked functional Notch1 and gave the same number of E9.5 embryos as eggs fertilized with a Notch1⁺ sperm.

Mutant embryos derived from Notch1^lbd/lbd eggs and therefore lacking maternal and zygotic transcripts of functional Notch1 were examined at E8.75 and E9.5. All Notch1^lbd/lbd E9.5 embryos were surrounded by a yolk sac with defective vascularization (Figs. 4B, C). Notch1^+/lbd and Notch1^lbd/lbd embryos at E8.75 (8 vs. 6 from 2 litters) were morphologically indistinguishable (Figs. 4D, E). By E9.5, the Notch1^lbd/lbd embryos were significantly smaller than controls, and the Notch1 null mutant phenotype was readily apparent (Figs. 4F, G). All developmental defects in Notch1^lbd/lbd embryos arising from mutant blastocysts with or without maternal Notch1 transcripts appeared similar. In addition, mutant embryos were morphologically similar to controls at E8.75 (Fig. 4D, E). Therefore Notch1 signaling induced by canonical Notch ligands is not required for oogenesis, ovulation, fertilization or any of the developmental steps involved in blastogenesis, implantation or gastrulation.

To generate the Notch1 ligand binding domain deletion mutation Notch1^lbd, an 1.6 kb region of genomic DNA containing exons 6 – 8 and two flanking sequences ~4.3 kb 5' (upstream) and ~2.9 kb 3' (downstream) were obtained by PCR from genomic DNA of WW6 ES cells 38 and cloned separately into pCR2.1 (T-vector, Invitrogen, Carlsbad, CA). A point mutation termed 12f was introduced into exon 8 by changing Thr466 to Ala 13. The integrity of the three inserts was confirmed by DNA sequencing and they were subcloned between the three loxP sites in the pFlox vector 39 using BamHI, SalI and XhoI with HindIII, respectively (Fig. 1A). The targeting vector was linearized using PvuI (the PvuI site between XhoI and HindIII in pFlox had been removed during subcloning). After gel purification, the plasmid was electroporated into WW6 ES cells using a Bio-Rad Gene Pulser (Bio-Rad, Hercules, CA) at 400 V and 250 μF. Following selection with 250 μg/ml active G418 (Invitrogen, Carlsbad, CA), resistant colonies were screened for homologous recombination by PCR using Takara Ex Taq (Takara Mirus Bio, Madison, WI) and primers N1ES-gF: 5'-GCTTCCCGCCTCCACTGTGCTATTGATGTTTG-3' from upstream of the 5' insertion site and pFlx-382R: 5'-GTTCCTCTTGCTGAACCACACTGCTCGATATTG-3' from the pFlox vector, and confirmed using primers pFlx-3521F: 5'-CTGTGCCTTCTAGTTGCCAGCCATCTGTTG-3' from the pFlox vector and DM142: 5'-CTGAAGCCTTCTCGGCAGGTGCATACGTAG-3' from downstream of the 3' insertion. Two positive ES clones were further characterized by Southern blot analysis after digestion by HindIII or HindIII and EcoRI. The probe P1415 is a genomic DNA fragment obtained by PCR from exons 14 to 15 of Notch1 using primers N1-ex14F: 5'-GTACAAGTGACTGTGCCCCTGGGTG-3' and N1-ex15R: 5'- CTGTATATGGCAGAGGACAGTTGCACTTG-3' and was used to determine integration into the endogenous Notch1 locus. Targeted ES cells were microinjected into C57Bl/6 blastocysts to obtain chimeric mice. Chimeras were crossed with transgenic mice carrying Cre recombinase under the control of a weak CMV promoter termed MeuCre40 14 to obtain heterozygous floxed Notch1 (Notch1^+/F) mice. Oocyte-specific Notch1 deletion was obtained by crossing Notch1^+/F mice with transgenic mice carrying Cre recombinase under the control of the ZP3 promoter (ZP3Cre) as previously described 2039. The deletion mutation was confirmed by Southern blot analysis after BamHI digestion using probe P1415 downstream of the 3' insertion site. Genotyping was performed from tail DNA or yolk sac DNA using forward primer 5F: 5'-GTATGTATATGGGACTTGTAGGCAG-3', and reverse primer 6R: 5'-CTATGAGGGGTCACAGGACCAT-3' that generate a 363 bp product from the wild type Notch1 allele and a 466 bp product from the floxed Notch1 allele, or primers 5F and 9R: 5'-CTTCATAACCTGTGGACGGGAG-3' that generate a 575 bp product from the Notch1 ligand binding domain deletion allele. ZP3Cre transgenic mice were genotyped as described 39.

Embryonic stem (ES) cell lines C1 (Notch1^+/+), C2 (Notch1^+/lbd) and A2 (Notch1^lbd/lbd) were isolated from E3.5 blastocysts obtained from Notch^+/lbd X Notch^+/lbd crosses as described 40 and genotyped by PCR from yolk sac DNA. Other ES cell lines used were WW6 38 (Notch1^+/+), and 290-2 Notch1 null cells 10 termed Notch1^in32/in32 and kindly provided by Dr. G. D. Longmore. ES cells were routinely cultured on a STO SNL2 feeder cell layer 41 in ES medium (Knockout-DMEM supplemented with 15% fetal bovine serum; Gemini, West Sacramento, CA, 1 × non-essential amino acids, 1 × L-glutamine, 1000 U/ml ESGRO^® (Chemicon, Temecula, CA), 50 μM ί-mercaptoethanol, 25 mM HEPES, penicillin (50 U/ml) and streptomycin (50 μg/ml). All reagents were ES-qualified and from SpecialtyMedia (Phillipsburg, NJ) except where mentioned). To remove feeder cells, ES cells were passaged on gelatinized plates for 2 – 3 generations at an 1:10 ratio. For growth curves, ES cells were plated on 24-well plates at 5 × 10⁴ cells per well and incubated at 37°C in an incubater with 5% CO₂. Cells from triplicate wells were trypsinized and counted after 24, 48, 72 and 96 h using a Z1 Coulter particle counter (Beckman Coulter, Fullerton, CA). Cells from each well were counted 3 times.

Total RNA was isolated from ES cells using TRIZOL^® (Invitrogen, Carlsbad, CA) followed by DNase I (Promega, Madison, WI) digestion according to the manufacturer's instructions. cDNA was prepared using the Takara RNA PCR Kit ver 3.0 (Takara Mirus Bio, Madison, WI). RT-PCR analysis was performed using Notch1 forward primer: 5'-GCCCTTTGAGTCTTCATACATCTG-3' and reverse primer: 5'-GACATTGGAACTCATTGATCTTGT-3'. PCR products (500 bp from Notch1^lbd and 1076 bp from Notch1⁺) were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. Gapdh was used as control (forward: 5'-ACCACAGTCCATGCCATCAC-3'; reverse: 5'-TCCACCACCCTGTTGCTGTA-3', product size: 452 bp).

ES cells growing on gelatinized plates were washed with PBS and lysed in RIPA buffer (Upstate, Lake Placid, NY) containing the complete protease inhibitors 'cocktail' (Roche, Basel, Switzerland). After incubation for 30 min on ice, the lysate was microfuged and protein concentration of the supernate was determined by Bio-Rad D_c protein assay (Bio-Rad, Hercules, CA). Lysates were resolved by 4–20% gradient SDS-PAGE, transferred to polyvinyldifluoride (PVDF) membranes and probed with Notch1 ECD antibody 8G10 (1:500; Upstate, Lake Placid, NY) to detect full length Notch1, or Notch1 antibody Val1744 (1:1000; Cell Signaling Technology, Beverly, MA) to detect activated Notch1. Horseradish peroxidase (HRP)-conjugated secondary antibodies were used to detect reactive bands visualized using Enhanced Chemiluminescence Reagent (Amersham Pharmacia Biotech, Piscataway, NJ). A β-Tubulin-III specific antibody (1:500; Sigma, St. Louis, MO) was used for loading control.

For cell surface Notch1 detection, ES cells growing on gelatinized plates at 70–80% confluence were dissociated using phosphate buffered saline (PBS) -based enzyme-free dissociation solution (SpecialtyMedia, Lavellette, NJ) for 10 min at 37°C. After washing with 10 ml medium, 5 × 10⁵ ES cells were incubated with 1 μg anti-Notch1 antibody (8G10, Upstate, Lake Placid, NY) in 100 μl FACS binding buffer (Hank's balanced salt solution (HBSS) containing 3% BSA, 0.05% sodium azide, and 1 mM Ca²⁺) for 1 h at 4°C in the dark, followed after washing with 1 ml FACS binding buffer, by incubation with 1:100 Alexa488-conjugated anti-hamster IgG antibody (Invitrogen, Carlsbad, CA) in 100 μl FACS binding buffer. After washing with 1 ml FACS binding buffer, the cells were suspended into 400 μl FACS binding buffer. Dead cells were excluded by staining with 7-AAD (BD Pharmingen, San Diego, CA). Flow cytometry was performed on a FACS Calibur flow cytometer (BD Biosciences, San Diego, CA). Data files were analyzed using Flowjo software (Tree Star, San Carlos, CA).

Soluble Notch ligand Delta1 with a human Fc tag (Delta1-Fc) 4243 was kindly provided by Dr. Gerry Weinmaster. HEK-293T cells expressing Delta1-Fc were cultured in αMEM (Invitrogen) containing 10% fetal bovine serum (Gemini). At 70–80% confluence, the medium was changed to 293 SFM II serum-free medium (Invitrogen) and culturing was continued. After 3 days, conditioned medium was collected, cellular debris removed by centrifugation, and the supernatant stored at 4°C. The concentration of ligand was determined by comparison with known concentrations of human IgG antibody (Jackson Immunoresearch, West Grove, PA) detected by chemiluminescence (Amersham Pharmacia Biotech, Piscataway, NJ) after western blotting. For the binding assay, 70–80% confluent ES cells were dissociated from plates using PBS-based enzyme-free dissociation solution (SpecialtyMedia, Lavellette, NJ) for 10 min at 37°C. After washing with 10 ml medium, the single cell suspension of ES cells (5 × 10⁵ cells) was incubated with 2 μg/ml Delta1-Fc in binding buffer (HBSS containing 3% BSA, 0.05% sodium azide, and 1 mM Ca²⁺) for 1 h at 4°C, followed after washing by incubation with 1:100 phycoerythrin (PE)-conjugated anti-human Fc antibody (Jackson Immunoresearch, West Grove, PA) for 30 min at 4°C. To inhibit Delta1 binding, 5 mM EDTA was added to the binding buffer. Flow cytometry was performed on a FACS Calibur flow cytometer. Ligand binding ability was determined by mean fluorescence intensity (MFI) of primary and secondary antibody binding minus the MFI of secondary antibody binding alone.

The co-culture Notch signaling assay was performed essentially as previously described 20. In brief, duplicate cultures of ES cells (Notch1^+/+,(C1) and Notch1^lbd/lbd (A2)) on 6-well plates were co-transfected with the TP1-luciferase Notch reporter plasmid and a Renilla luciferase reporter (pRL-TK; Promega, Madison, WI) and empty vector using FuGENE 6 (Roche, Basel, Switzerland). At 16 h post-transfection, ES cells were overlaid with 1 × 10⁶ Jagged1-expressing L cells or Delta1-expressing L cells or parental L cells. At ~40 h post-transfection, firefly and Renilla luciferase activities were quantitated in cell lysates using a dual luciferase assay kit (Promega, Madison, WI) on Autolumat Plus LB 953 (Berthold Technologies, Bad Wildbad, Germany) according to the manufacturer's instructions. Ligand-dependent Notch activation is expressed as fold-induction of normalized firefly luciferase activity obtained from Notch ligand versus L cell co-cultures. Co-culture assays with Notch1^in32/in32 ES cells (290-2) compared to Notch1^+/+ ES cells (C1) were performed by the same method in 12-well plates using Lipofectamine (Invitrogen) for transfection of plasmids.

Embryos of E8.75, E9.0 or E9.5 from Notch1^+/lbd crosses were harvested and DNA was prepared from yolk sac for genotyping. Embryos for whole mount in situ hybridization were fixed in 4% formaldehyde in PBS overnight at 4°C. Whole-mount in situ hybridization was performed as previously described 16. The hybridization probes used were: Notch1 full length, ~9.5 kb or ~4.7 kb 44; Uncx4.1 ~1.7 kb 45, Hes5 ~1.3 kb 46; and Myogenin ~1.5 kb 47. Stained embryos were photographed in PBS through a phototube on Leica Wild M3Z stereomicroscope (Leica-Microsystems, Heerburgg, Switzerland) using a Canon S40 digital camera (Canon USA Inc., Lake Success, NY).