Expression pattern of Motch, a mouse homolog of Drosophila Notch, suggests an important role in early postimplantation mouse development

The differentiation of the nervous system is a key event in the early development of many organisms. In Drosophila, the central nervous system originates from the ventrolateral neurogenic region of the ectoderm (for reviews, see Campos-Ortega and Hartenstein, 1985; Campos-Ortega and Jan, 1991). In wild-type embryos, the neuroblasts, the progenitor cells of the embryonic nervous system, delaminate from the epithelial layer, move interiorly and divide to produce the neurons of the brain and the ventral nerve cord. Cells remaining in the epithelial layer of the neurogenic region adopt a different developmental fate. These cells develop as epidermoblasts, which give rise to ventral epidermis and secrete the embryonic cuticle.

This choice of cell fate in the neurogenic region is under genetic control (Artavanis-Tsakonas et al., 1991; Campos-Ortega and Jan, 1991; Campos-Ortega and Knust, 1990). In embryos homozygous for null mutations of the Notch gene, essentially all of the cells in the neurogenic region become neuroblasts (Lehman et al., 1983; Poulson, 1937). Such embryos die with a vast hypertrophy of the nervous system and a corresponding absence of epidermal structures. Embryos homozygous for null mutations of at least five other zygotically acting loci (Delta, Enhancer of split, mastermind, big brain and neuralized) exhibit a similar mutant phenotype (Lehman et al., 1983). Cell ablation studies during neurogenesis in grasshopper embryos, which have a mode of neural development similar to Drosophila embryos, have indicated that cell-cell interactions are important in the decision of a bipotential precursor cell to develop as either a neuroblast or an epidermoblast (Doe and Goodman, 1985).

Genetic and biochemical analyses of the Notch gene have suggested that Notch could play a role in the cell-cell interactions important in determining cell fate in the neurogenic region (Fehon et al., 1990, 1991; Kidd et al., 1989; Rebay et al., 1991; Xu et al., 1990; for reviews, see Artavanis-Tsakonas and Simpson, 1991; Artavanis-Tsakonas et al., 1991; Campos-Ortega and Jan, 1991; Greenspan, 1990; Greenwald and Rubin, 1992). The Notch gene encodes a large transmembrane protein whose extracellular domain contains 36 tandemly repeated copies of an epidermal growth factor (EGF)-like sequence (Kidd et al., 1986; Wharton et al., 1985). EGF-like sequences have been shown to act as sites for protein-protein interactions (Apella et al., 1987; Kurosawa et al., 1988), and these sequences are present in a wide variety of proteins (Bevilacqua et al., 1989; Jones et al.,1988; Lawler and Hynes, 1986; Mann et al., 1989; Montell and Goodman, 1988; Siegelman et al., 1989; Sudhof et al., 1985; Suzuki et al., 1987; for a review, see Davis, 1990). The extracellular domain also contains three repeats of another cysteine-rich motif, termed the Notch/lin-12 repeat. This motif is found in the Notch gene as well as in two genes in Caenorhabditis elegans, lin-12 and glp-1 (Yochem and Greenwald, 1989). These two genes, which encode transmembrane proteins which also contain EGF-like repeats in their extracellular domain, are involved in cell-cell interactions and cell fate specification during nematode development (for reviews, see Greenwald and Rubin, 1992; Maine and Kimble, 1990). The intracellular domains of Notch, as well as lin-12 and glp-1, also contain six copies of another motif termed the cdc10/ankyrin repeats. These repeats have been identified on an increasingly broad range of proteins (Aves et al., 1985; Breeden and Nasmyth, 1987; Davis and Bennett, 1990; Lux et al., 1990; Ohno et al., 1990; Spence et al., 1990; Thompson et al., 1991), and have recently been shown to be domains involved in protein-protein interaction (Davis and Bennett, 1990; Thompson et al., 1991).

The isolation of murine homologs of genes important for development in other organisms has been an important addition to the array of techniques that can be used to study mammalian development (Chisaka and Capecchi, 1991; Joyner et al., 1991; McMahon and Bradley, 1990; for reviews, see Gridley, 1991; Kessel and Gruss, 1990). To determine if a gene homologous to Drosophila Notch plays a role in early mouse development, we have screened an 8.5 dpc mouse embryo cDNA library with probes from the Xeno - pus Notch homolog, Xotch (Coffman et al., 1990). We report here the isolation of a partial cDNA clone encoding the mouse homolog of Notch, which we have termed Motch. We characterize by ribonuclease protection analysis the expression pattern of Motch RNA in adult mice and postimplantation mouse embryos, and the induction of Motch RNA levels in tissue culture cells undergoing differentiation. An analysis by in situ hybridization of the spatial organization of Motch RNA expression suggests multiple roles for the Motch gene in patterning and differentiation in the early postimplantation mouse embryo.

To isolate clones that cross-hybridize with the Xotch cDNA (Coffman et al., 1990), an 8.5 dpc mouse embryo cDNA library (Fahrner et al., 1987) and a mouse genomic library (Kinloch et al., 1988) were screened with two Xotch cDNA probes at low stringency. Plaque lifts were hybridized in 5 × SSC, 50 mM Tris-HCl, pH 8.0, 2.5 × Denhardt’s, 100 μg ml⁻¹ yeast RNA, 10% dextran sulphate, 1.0% SDS, 1.0% sodium pyrophosphate at 55°C, and final wash conditions were 1 × SSC, 0.2% SDS at 55°C. The probes were a 900 bp EcoRV-EcoRV fragment encoding amino acids 1111–1445 of the predicted Xotch protein (which includes part of the EGF-like repeat region) (Coffman et al., 1990), and a 500 bp EcoRV-Bgl II fragment encoding amino acids 1446-1612 (which includes the Notch/lin-12 repeat region). One cDNA (c195) was isolated that hybridized with the Notch/lin-12 repeat probe, and one genomic (g1) phage was isolated that hybridized with both Xotch probes. DNA fragments isolated from recombinant phage carrying cDNA or genomic sequences were subcloned into the plasmid vector pGem7 (Promega). DNA sequencing was performed by the dideoxy technique using the Sequenase enzyme (US Biochemical Corporation).

Total RNA was isolated by the acid guanidinium thiocyanate-phenol-chloroform technique (Chomczynski and Sacchi, 1987) using RNazol B (Biotecx). Ribonuclease protection analysis was performed with the RPA II Ribonuclease Protection Assay kit (Ambion), using ³²P-labelled antisense strand of the EcoR1-Kpn1 fragment of the Motch cDNA clone c195 as the antisense riboprobe. Quantitation of ribonuclease protection results was performed on a Betascope 603 blot analyzer (Betagen). The Motch signal (counts per minute) was normalized to the β-actin signal, and the highest normalized signal was arbitrarily set at 100 units. Results were similar if the Motch signal was normalized to a second housekeeping gene, the product of the Mov-34 locus (Gridley et al., 1990, 1991). The results shown are representative of an experiment repeated four times.

Fig. 3. Motch RNA expression in embryos at 6.5 and 7.5 dpc. All sections were hybridized with a Motch antisense riboprobe. (A) Sagittal section of 6.5 dpc embryo. No Motch expression is observed in any embryonic or extra-embryonic region of the embryo. Motch is expressed at this time in maternally derived cells, primarily at the mesometrial pole of the decidua (not shown). (B) A slightly oblique sagittal section of 7.5 dpc embryo. Motch expression is observed in mesoderm and in embryonic ectoderm adjacent to the primitive streak. (C) Transverse section of 7.5 dpc embryo. Expression is again observed in mesoderm and embryonic ectoderm. Motch expression in the mesoderm appears to be graded. The highest levels are in the primitive streak, and they appear to decline in more lateral and anterior mesoderm (compare B). The open arrows delimit the anterior limits of extension of the mesodermal wings. Expression in the decidua is more widespread than at 6.5 dpc, and is now observed in maternally derived cells surrounding the embryo. am, amnion; e, embryonic ectoderm; ec, ectoplacental cone; ee, extra-embryonic ectoderm; m, mesoderm; ps, primitive streak.

Fig. 5. Motch RNA expression in embryos at 9.5 and 10.5 dpc. All sections were hybridized with a Motch antisense riboprobe. (A-C) Sections of a 9.5 dpc embryo. (D-H) Sections of a 10.5 dpc embryo. (A) Very high levels of expression are observed in presomitic mesoderm. Expression is also observed in the neural tube and in endothelial cells lining the dorsal aorta. (B) Motch is widely expressed in the developing brain, and also appears to be expressed in the meninges (arrow). (C) Motch is expressed in endocardium of the heart and in endothelial cells lining blood vessels in the branchial arches (arrows). (D) In this section Motch expression is observed in the spinal cord, in the condensing dorsal root ganglia, in mesonephric tubules, and in intersomitic blood vessels (arrows). (E) Higher magnification view of the section shown in (D). Expression is observed in endothelial cells lining the intersomitic blood vessels (arrows) and in the sclerotome. No expression is observed in dermatome or myotome. (F) Motch expression is observed in the mitotic cells of the ventricular zone of the spinal cord, and in the condensing dorsal root ganglia. No expression is observed in the floor plate or in postmitotic cells of the ventral horn of the spinal cord. (G) Motch is expressed in the trigeminal ganglion. (H) High levels of expression are observed in surface ectoderm of the lateral nasal process. Lower levels of expression are observed in ectoderm and mesenchyme of the medial nasal process. d, dermatome; da, dorsal aorta; dr, condensing dorsal root ganglion; e, endocardium; fp, floor plate; lnp, lateral nasal process; m, myotome; mes, mesencephalon; mnp, medial nasal process; ms, mesonephric tubules; op, olfactory pit; s, scleratome; tel, telencephalon; tg, trigeminal ganglion; vh, ventral horn of the spinal cord; vz, ventricular zone.

Fig. 6. Motch expression in 13.5 dpc embryos. All sections were hybridized with a Motch antisense riboprobe. (A) Parasagittal section of the nasal region of an embryo at 13.5 dpc. Expression is observed in epidermis and developing whisker follicles. At this stage of gestation, Motch appears to be expressed in all surface ectoderm. (B) Higher magnification view of the section shown in A. Motch expression in the developing whisker follicle is confined to the inner root sheath. (C) Sagittal section through the eye of an embryo at 13.5 dpc. Motch expression is observed in the sensory layer of the retina and in the lens. Apparent expression in the pigment layer of the retina (asterisk) is an artifact due to light scattering by the pigment granules. (D) Higher magnification view of the section shown in C. Motch expression is not observed in differentiating cells of the ganglion cell layer of the retina. Motch expression is observed in mitotic cells in the sensory layer of the retina and in epithelial cells of the presumptive cornea. *, pigment layer of the retina; e, epidermis; gc, ganglion cell layer of the retina; ir, inner root sheath; l, lens; pc, presumptive cornea; r, retina; wf, whisker follicle.

P19 embryonal carcinoma cells were cultured and induced to differentiate essentially as described (Rudnicki and McBurney, 1987). Cells were grown in αMEM plus 10% fetal bovine serum. To induce differentiation, cells were placed in suspension culture in agarose-coated dishes, either as a single cell suspension or as lightly trypsinized aggregates of cells (as described for embryonic stem cells by Robertson, 1987). Cells placed in suspension culture were either in αMEM plus 10% fetal bovine serum (untreated controls) or in the same media plus 0.3 μM retinoic acid (Sigma). After four days in suspension, all samples contained aggregates of P19 cells. These aggregates were then plated out in 10 cm tissue culture dishes in media without retinoic acid. Microscopic examination of these cultures after five days revealed the presence of cells containing neuronal processes in retinoic acid-induced cultures, but not in untreated control cultures. RNA was harvested from these cells using RNAzol B, and RNA samples were assayed for Motch and β-actin transcripts by ribonuclease protection.

In situ hybridization was performed essentially as described by Wilkinson and Green (1990). Briefly, C57Bl/6 embryos were dissected and fixed at 4°C in 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight. The fixed embryos were dehydrated, embedded in paraffin, and 6 μm sections were cut and floated onto 3-aminopropyltriethoxysilane-coated slides. For hybridization, slides were dewaxed in xylenes, hydrated in an ethanol series and fixed in fresh 4% paraformaldehyde in PBS. Sections were treated with 20 μg ml⁻¹ proteinase K in 50 mM Tris-HCl, 5 mM EDTA, pH 8.0, washed in PBS and postfixed in 4% paraformaldehyde in PBS. Sections were then treated with acetic anhydride, washed and dehydrated. [³⁵S]UTP-labelled singlestranded sense and antisense RNA probes were prepared by standard procedures (Sambrook et al.,1989). The probe was hydrolyzed to an average length of 100 bases, unincorporated nucleotides were removed by chromatography on a Nick column (Pharmacia), and the probe was ethanol-precipitated. The probe was resuspended at a concentration of 2 ng μl⁻¹ kb⁻¹ in 100 mM DTT. The probe was then diluted 1:10 in hybridization solution (50% formamide, 0.3 M NaCl, 20 mM Tris-HCl, 5 mM EDTA, pH 8.0, 10% dextran sulphate, 1 × Denhardt’s, 0.5 mg ml⁻¹ yeast RNA), giving a final probe concentration of 0.2 ng μl⁻¹ kb⁻¹. After sections were hybridized overnight at 55°C, they were treated with ribonuclease A, washed at high stringency (50% formamide, 2 × SSC, 10 mM DTT at 65°C) and dehydrated. Slides were dipped in NTB2 emulsion (Eastman Kodak) and exposed for approximately two weeks. Exposed slides were developed in D19 (Kodak), stained in 0.5% toluidine blue and mounted with Permount (Fisher). Double exposure (dark-field with red filter, bright-field with blue filter) photomicrographs were taken on a Zeiss Axioplan microscope.

To determine if a gene homologous to Drosophila Notch plays a role in early mouse development, we screened mouse embryo cDNA (Fahrner et al., 1987) and genomic libraries at low stringency with cDNA probes encoding two domains of Xotch, the Xenopus Notch homolog (Coffman et al., 1990). One Xotch probe encoded part of the EGF-like repeat region of Xotch, while the other probe encoded the Notch/lin-12 repeat region. This screen yielded a cDNA clone that hybridized to the Xotch probe encoding the Notch/lin-12 repeat region. The 2.0 kb insert from this cDNA clone, c195, was subcloned and sequenced. Fig. 1A shows the nucleotide and translated amino acid sequences for the 291 bp EcoR1 - Kpn1 fragment at the 5′ end of this cDNA clone. After the initiation of these studies, the cloning of Notch homologs from human (Ellisen et al., 1991) and rat (Weinmaster et al., 1991) was reported. Comparison of the sequence of the mouse cDNA clone with the sequences of Notch and the other vertebrate Notch homologs confirms that this clone encodes the mouse Notch homolog, which we have named Motch. This EcoR1 - Kpn1 fragment encodes the last 13 amino acids of the third Notch/lin-12 repeat, as well as the succeeding 84 amino acids of the Motch gene. Fig. 1B compares the translated amino acid sequence of this subclone with the amino acid sequences for Notch and the other published Notch homologs cloned from Xenopus, rat and man. The last line in Fig. 1B displays the consensus sequence for this region of the Notch family of proteins. In this region approximately 26% (26/98) of the amino acids are identical in all the Notch homologs, while 56% (55/98) of the amino acids are identical in all the vertebrate Notch homologs. Comparing the mouse and rat Notch homologs, 95% (92/97) of the amino acids in this region are identical. The EcoR1 - Kpn1 fragment of the Motch gene described above was subcloned and used to analyze Motch RNA expression by ribonuclease protection and in situ hybridization. Cloning of a full-length cDNA of the Motch transcript is in progress and will be described in a separate report. In addition, screening of a mouse genomic phage library (Kinloch et al., 1988) with the two Xotch probes also yielded a single clone which hybridized to both Xotch probes. Preliminary nucleotide sequence analysis of this genomic clone indicates that it contains part of the Motch gene encoding some of the EGF-like repeats as well as the Notch/lin-12 repeats (data not shown).

Motch RNA expression in adult mouse tissues was analyzed utilizing a ribonuclease protection assay, using the cDNA fragment shown in Figure 1A as an antisense riboprobe. Motch transcripts were detected in a wide variety of adult tissues and included derivatives of all three germ layers (Fig. 2A). Motch RNA levels were highest in brain, lung and thymus. Lower levels of RNA were detected in spleen, bone marrow, spinal cord, eyes, mammary gland, liver, intestine, skeletal muscle (not shown) and heart. These data agree well with the expression pattern reported for 10-20 week old human fetuses analyzed for expression of the human Notch homolog, TAN-1 (Ellisen et al., 1991). Motch RNA expression during embryonic development was also examined with the ribonuclease protection assay. Motch RNA was present during all post-implantation stages examined (from 7.5 dpcto 16.5 dpc) (Fig. 2B).

Fig. 4. Motch expression in 8.5 dpc embryos. All sections were hybridized with a Motch antisense riboprobe. (A) Schematic diagram of a mouse embryo at 8.5 dpc. The approximate levels of sections B-G are indicated. The extent of the foregut diverticulum is indicated by the dotted line. (B-F) Transverse sections. Highest levels of Motch expression are observed in presomitic mesoderm (B-E). Motch expression is substantially down-regulated when the paraxial mesoderm condenses to form somites (F). Motch transcripts are also abundant in mesenchyme (B-E) and endocardium (D,E). (G) Slightly oblique frontal section. Motch expression is observed in cephalic mesenchyme, but little or no expression is observed in neuroepithelium. cm, cephalic mesenchyme; cs, condensed somite; e, endocardial cells forming inner tube of the primitive heart; f, foregut; pm, presomitic mesoderm.

We were interested in determining if Motch RNA levels were altered in cell culture systems that can undergo differentiation. Suspension culture of P19 embryonal carcinoma cells in the presence of retinoic acid leads to the differentiation of neuronal and glial cell types upon subsequent plating of the cells (Rudnicki and McBurney, 1987). Motch RNA levels were determined by ribonuclease protection before and after induction of differentiation of P19 cells. The Motch gene is expressed in undifferentiated P19 cells. Motch RNA levels (normalized to RNA levels for β-actin - see Materials and methods) increased approximately 8-fold after induction with retinoic acid (Fig. 2C,D). The increase in Motch RNA levels was similar if the signal was normalized to a second housekeeping gene, the product of the Mov-34 locus (Gridley et al., 1990, 1991). Motch RNA induction was consistently greater if the P19 cells were initially placed in suspension culture as aggregates, rather than as a single cell suspension.

To further characterize Motch RNA expression during postimplantation development, we analyzed the spatial distribution of Motch transcripts by in situ hybridization. Since we were interested in determining if the Motch gene might play similar roles in early embryonic development of Drosophila and mice (e.g., cell fate determination in the developing nervous system), we concentrated our analysis on early postimplantation stages (6.5 dpc to 10.5 dpc). No Motch transcripts could be detected in any embryonic tissue at 6.5 dpc (Fig. 3A), although Motch transcripts were detected in maternally derived cells at the mesometrial pole of the decidua at this time (data not shown). At 7.5 dpc, Motch transcripts were first detected in mesoderm and in posterior embryonic ectoderm (Fig. 3B,C), in and adjacent to the primitive streak. No transcripts were detected in any extraembryonic tissues (visceral endoderm, amnion, chorion, ectoplacental cone) at this time.

At 8.5 dpc the highest level of Motch expression was in presomitic mesoderm (Fig. 4B-E). Motch expression was down-regulated when mesoderm condensed to form somites (Fig. 4F). Motch transcripts were also abundant in cephalic mesenchyme (Fig. 4D,E) and in endothelial cells lining the inner heart tube and the dorsal aorta (Fig. 4D,E). Much lower levels of Motch transcripts were present in neuroepithelium. Expression in neuroepithelium at 8.5 dpc appeared to be at about the limit of detection in our in situ hybridization experiments.

By 9.5 dpc, however, neuroepithelial tissues were a major site of Motch expression, and transcripts were detected in many areas of the brain and neural tube (Fig. 5A,B). Motch RNA levels remained high in presomitic mesoderm (Fig. 5A). We continued to detect Motch transcripts in endothelial cells at 9.5 and 10.5 dpc; transcripts could be observed in endocardium (Fig. 5C) and in endothelial cells lining the dorsal aorta (Fig. 5A), the intersomitic blood vessels (Fig. 5D,E) and the aortic branches in the branchial arches (Fig. 5C). At 10.5 dpc we continued to detect Motch transcripts in neuroepithelium. Motch expression in the spinal cord was confined to mitotic cells of the ventricular zone; postmitotic cells located in the ventral horns did not express (Fig. 5D,F). We also detected transcripts in mesonephric tubules (Fig. 5D), condensing dorsal root ganglia (Fig. 5D,F), the trigeminal ganglion (Fig. 5G) and the lateral nasal process (Fig. 5H). In a less extensive analysis of embryos at 13.5 dpc, we detected expression in surface ectoderm, in the eye and in the developing whisker follicles (Fig. 6). Expression in the whisker follicle was confined to mitotic cells of the inner root sheath (Fig. 6B). Similarly, Motch expression in the eye was confined to mitotic cells of the lens, corneal epithelium and sensory layer of the retina. Differentiating ganglion cells of the neural retina did not express Motch (Fig. 6D).

We report here an analysis of the RNA expression pattern of Motch, a mouse homolog of Drosophila Notch. After the initiation of these studies, the cloning of Notch homologs from human (Ellisen et al., 1991) and rat (Weinmaster et al., 1991) was reported. The vertebrate Notch gene family also includes Xotch, the Xenopus Notch homolog (Coffman et al., 1990) and a second mouse gene, the product of the int-3 locus (Gallahan et al., 1987), which is more distantly related to Notch than the genes mentioned above (Robbins et al., 1992). The vertebrate Notch genes are very well conserved and appear to be true homologs of Notch : all the apparent structural motifs of Notch (the EGF-like repeats, the Notch/lin-12 repeats and the cdc10/ankyrin repeats) are conserved, both in the number of repeats present and the location and order of these structural motifs in the protein. Even a skeleton of the opa repeat, the polyglutamine stretch present in a number of Drosophila genes, is retained. Thus, it seems quite likely that the biochemical mode of action will be similar in all the Notch homologs. Given the strong structural conservation of the Notch family of genes, we were interested in determining if the mode of action of these proteins was functionally conserved as well. In particular, we wanted to determine if Motch, the mouse Notch homolog, was involved in early differentiative events of the developing mouse embryo, e.g. formation of the mesoderm or the neural plate. As a first step to answering this question, we examined by in situ hybridization the RNA expression pattern of the Motch gene during early postimplantation development (6.5 – 10.5 dpc).

Our results showed that, in postimplantation embryos, Motch transcripts were first detected at 7.5 dpc in mesoderm and posterior embryonic ectoderm, in and adjacent to the primitive streak. This region of the embryonic ectoderm consists of epithelial cells destined to delaminate and form mesoderm and definitive endoderm (Beddington, 1981; Beddington, 1982; Tam and Beddington, 1987). Motch transcripts at 8.5 dpc were most abundant in presomitic mesoderm, cephalic mesenchyme and endothelial cells of the endocardium and dorsal aorta, but were barely detectable in neuroepithelium itself at this stage. A day later, however, neuroepithelial tissues have become one of the major sites of Motch expression in the embryo. Thus, Motch expression may not be required in the cells of the neural plate for their specification and initial differentiation. However, Motch is expressed in neuroepithelium after it first forms.

In addition to neuroepithelium of the central nervous system, we also observed Motch expression in a number of other sites. Several of these sites are derived from neural crest cells. Motch expression was observed in the condensing dorsal root ganglion at 10.5 dpc. Other sites of Motch expression, such as the trigeminal ganglion and the cephalic mesenchyme, are also derived at least in part from neural crest cells. As mentioned previously, another site of Motch expression was endothelium, which is of mesodermal origin. Our results support and extend the expression data published for the other vertebrate Notch homologs. Coffman et al. (1990) found, using a ribonuclease protection assay, that Xotch, the Xenopus homolog, was expressed almost uniformly in early embryos and was present in all three prospective germ layers. Weinmaster et al. (1991) used the rat Notch homolog to examine expression in mouse embryos from 9.5 to 16.5 dpc by in situ hybridization. They observed strong expression in developing brain, spinal cord, eyes, dorsal root ganglia and trigeminal ganglia between 10 and 12 dpc. Expression in later embryos was observed in tissues undergoing epithelial-mesenchymal interactions, such as whisker follicles, tooth buds and kidney (Weinmaster et al., 1991). We also observed expression in eyes and whisker follicles at 13.5 dpc.

We further observed that Motch transcript levels increase when P19 embryonal carcinoma cells undergo differentiation into neuronal cell types. Since Motch RNA is expressed at low levels in undifferentiated P19 cells, it is not clear at present if this induction of Motch RNA levels has any functional significance. We plan further experiments to modulate Motch expression in P19 cells to determine if this has an effect on the ability of these cells to differentiate.

Our expression results suggest that Motch is playing multiple roles during development of mouse embryos and in the adult. This is consistent with the involvement of Notch in several developmental events in Drosophila besides neuroblast segregation (Cagan and Ready, 1989; Hartenstein and Posakony, 1990; Portin, 1975; Ruohola et al., 1991). Our results further suggest that the first site of action of the Motch gene during mouse embryogenesis is in mesoderm. This result is interesting in light of recent findings on the functioning of the neurogenic genes in mesoderm of the Drosophila embryo. Corbin et al. (1991) demonstrated that in addition to specifying cell fate in the neurogenic ectoderm, Notch and the other neurogenic genes also appear to be required for correct specification of cell fates in mesoderm of Drosophila embryos.

While it seems quite likely that the biochemical mode of action will be similar in all the Notch homologs, it is not entirely clear at present what role Notch actually plays in Drosophila development. Several models have been proposed for Notch function. One proposed role for Notch is that it functions as the receptor of the signal for lateral inhibition. Notch has been shown to function autonomously in both embryonic (Hoppe and Greenspan, 1990) and imaginal (de Celis et al., 1991; Heitzler and Simpson, 1991; Markopoulou and Artavanis-Tsakonas, 1991) development, as would be expected for a receptor. Autonomy has also been demonstrated for the Notch-related genes in C. elegans, lin-12 and glp-1 (Austin and Kimble, 1987; Seydoux and Greenwald, 1989). Most, but not all (Technau and Campos-Ortega, 1987; however, see Simpson, 1990), of the experimental results obtained to date support the model that Notch functions autonomously, probably as the receptor for a signal. This would not exclude, however, the possibility that Notch may have additional functions as well. The early expression pattern of Motch does not seem to support a role in binary cell fate decisions, although a mutational analysis of the Motch gene will be required to help answer this question.

Another function that has been proposed for Notch is a role in cell proliferation. Both in situ hybridization and immunohistochemistry experiments have demonstrated a good, although not absolute, correlation during Drosophila development between Notch expression and mitotically active cell populations (Kidd et al., 1989; Markopoulou and Artavanis-Tsakonas, 1989). Both our results and those of Coffman et al. (1990) and Weinmaster et al. (1991) indicate that the vertebrate Notch genes also show a correlation, particularly in the developing nervous system, between expression and cell proliferation. In addition, while the Xenopus, rat and mouse Notch homologs were all cloned by low-stringency hybridization techniques, the human Notch homolog, TAN-1, and the product of the mouse int-3 locus were identified as oncogenes. TAN-1 was identifed at a translocation breakpoint in a T cell lymphoma (Ellisen et al., 1991), while the int-3 locus is a common integration site for Mouse Mammary Tumor Virus (MMTV)-induced mammary tumors (Gallahan et al., 1987; Robbins et al., 1992). In both cases it appears that deregulated expression of the cytoplasmic domain, containing the cdc10/ankyrin repeats, is the likely cause of the tumorigenic phenotype. This has been particularly well established for the int-3 locus, since a minigene construct containing essentially only the cytoplasmic domain under the transcriptional control of the MMTV long terminal repeat can cause mammary tumors in transgenic mice (Jhappan et al., 1992). It is clear, then, that mutations of vertebrate Notch genes can affect cell proliferation. Whether the normal function of these genes involves a role in cell proliferation remains to be demonstrated.

Another model that has been proposed for Notch is that its expression is required to stabilize and/or maintain the differentiated state (Hoppe and Greenspan, 1990). It is possible that Motch expression may also be serving this function. For example, Motch would not be required for specification and initial formation of neuroepithelium, but would be required for its maintenance, hence the widespread expression of Motch in neuroepithelium at 9.5 dpc.

A fourth function that has been proposed for Notch is a role in cell adhesion (Cagan and Ready, 1989; Greenspan, 1990; Hoppe and Greenspan, 1990). The RNA expression pattern of Motch, however, gives no indication that it plays a role, at least initially, in cell adhesion. In a number of instances, Motch RNA is expressed in mesenchymal or migrating cells (e.g. invaginating mesoderm and cephalic mesenchyme). It will be important to analyze expression of Motch protein to determine if it is expressed concomitantly with the first appearance of Motch transcripts in mesodermal cells.

The expression pattern of the Motch gene suggests that it plays an important role in patterning and differentiation of early postimplantation mouse embryos. Proof of such a role, however, will require a functional analysis in transgenic mice. Analysis of both gain-of-function (i.e. ectopic expression) and loss-of-function (i.e. construction of null mutations in embryonic stem cells) mutations will greatly aid our understanding of Motch gene function in early mouse development.

We would like to thank Jill McMahon, Galya Vassileva, and Mary Dickinson for advice on in situ hybridization, Clark Coffman, Chris Kintner and Bill Harris for the Xotch probes, Ross Kinloch and Paul Wassarman for the mouse genomic phage library, and Joe Grippo for use of his microscope. F.F. del A. was supported by a fellowship from the Ministerio de Educación y Ciencia of Spain (#FP90-32450447).

The nucleotide sequence of the Motch cDNA clone c195 will appear in the EMBL/Genbank/DBJ databases under the accession number Z11886.