During neurogenesis in the ventral nerve cord of the Drosophila embryo, Notch signaling participates in the pathway that mediates asymmetric fate specification to daughters of secondary neuronal precursor cells. In the NB4-2 → GMC-1 → RP2/sib lineage, a well-studied neuronal lineage in the ventral nerve cord, Notch signaling specifies sib fate to one of the daughter cells of GMC-1. Notch mediates this process via Mastermind (Mam). Loss of function for mam, similar to loss of function for Notch, results in GMC-1 symmetrically dividing to generate two RP2 neurons. Loss of function for mam also results in a severe neurogenic phenotype. In this study, we have undertaken a functional analysis of the Mam protein. We show that while ectopic expression of a truncated Mam protein induces a dominant-negative neurogenic phenotype, it has no effect on asymmetric fate specification. This truncated Mam protein rescues the loss of asymmetric specification phenotype in mam in an allele-specific manner. We also show an interallelic complementation of loss-of-asymmetry defect. Our results suggest that Mam proteins might associate during the asymmetric specification of cell fates and that the N-terminal region of the protein plays a role in this process.
THE central nervous system (CNS) of the Drosophila embryo provides an important paradigm for investigating the problem of asymmetric division of neural precursor cells during development. In the ventral nerve cord of the Drosophila embryo, ∼30 neuroblast (NB) cells in each hemi-segment delaminate in about five successive waves along the mediolateral and anterior-posterior axes in rows and columns in a stereotyped and spatio-temporal pattern (Hartenstein and Campos-Ortega 1984; Doe 1992). Each of these NBs has acquired a unique fate by the time it is formed, and the NB that forms in a given position at a given time always acquires the same fate (reviewed in Bhat 1999). A neuroblast then functions as a stem cell and divides by asymmetric mitosis, renewing itself with each division and producing a chain of ganglion mother cells (GMCs). A GMC does not self-renew; instead it divides to generate two distinct neurons. These postmitotic neurons then undergo cyto-differentiation. At the end of neurogenesis, each of the hemi-neuromeres has ∼320 neurons and ∼30 glia, the other principal cell type in the CNS (Bossinget al. 1996; Schmidtet al. 1997). Thus, a complex array of different cell types is formed from relatively few precursor cells.
Genetic and molecular evidence indicate that GMCs generally undergo an asymmetric cell division (Hartenstein and Posakony 1990; Bhat and Schedl 1994; Bhatet al. 1995; Hirataet al. 1995; Knoblichet al. 1995; Spana and Doe 1995, 1996; Buescheret al. 1998; Dyeet al. 1998; Skeath and Doe 1998; Learet al. 1999; Waiet al. 1999; Mehta and Bhat 2001). One of the earliest evidences comes from a study on the development of the adult sensilla in which the neurogenic gene Notch plays a role in generating asymmetric division of secondary precursor cells (Hartenstein and Posakony 1990). Using the temperature-sensitive allele of Notch, it was shown that eliminating Notch activity in sensillum precursors leads to hyperplasia of the sensory neurons at the expense of accessory cells (i.e., shaft, socket cells). Notch, together with Numb (Nb), also regulates asymmetric fate specification to progeny of GMC in the ventral nerve cord (Buescheret al. 1998; Skeath and Doe 1998; Learet al. 1999; Waiet al. 1999). In the GMC-1 → RP2/sib lineage, loss of Notch or nb leads to the symmetric division of GMC-1. While both progeny assume RP2 fate in Notch mutants, they assume a sib fate in nb mutants (Buescheret al. 1998; Learet al. 1999; Schuldt and Brand 1999; Waiet al. 1999). Nb appears to block the intracellular domain of Notch from being cleaved and translocated to the nucleus, thus allowing that cell to adopt an RP2 fate. These studies have also revealed that the gene product of mastermind (mam) is downstream of Notch and that loss of function for mam results in both daughters of GMC-1 adopting an RP2 fate.
Mam is a glutamine-rich nuclear protein essential for Notch signaling (Smolleret al. 1990; Bettleret al. 1996; Helmset al. 1999). Mam interacts with the N intracellular domain (Nintra) in the Suppressor of Hairless [Su(H)]/CBF1 complex and stabilizes its binding to DNA in vitro (Petcherski and Kimble 2000a,b; Wuet al. 2000; Kitagawaet al. 2001). A recent study to evaluate the activity of the Notch minimal functional enhancer complex in a mammalian chromatin-based, cell-free transcription system indicates that human Mam is an essential component of this complex that associates with histone acetyltransferases (Fryeret al. 2002). Mam has one basic domain at the N terminus and two acidic domains, one close to the middle of the protein and another at the C terminus. The basic region of Mam is conserved in fly, mouse, and human, and this region physically interacts with the processed Nintra (Petcherski and Kimble 2000a; Wuet al. 2000; Kitagawaet al. 2001). Truncations in Mam that remove parts of the protein carboxy to the basic region elicit dominant-negative phenotypes when overexpressed in imaginal tissues (Helmset al. 1999). A further functional dissection of the Mam protein in vivo during neurogenesis has not been done. Given that loss of Mam activity leads to two distinct phenotypes in the Drosophila CNS—neural hyperplasia and loss of asymmetric division of secondary neuronal precursor cells—we attempted to distinguish these functions of the Mam protein in vivo. We examined transgenic lines expressing truncated versions of the Mam protein and determined whether they have a neurogenic and or a loss-of-asymmetry phenotype and whether any of the truncated proteins would rescue the asymmetry phenotype of mam. We show that while ectopic expression of a Mam truncation induces a dominant-negative neurogenic phenotype, it has no effect on the asymmetric fate specification to daughter cells of secondary neuronal precursor cells. Consistent with this result, this truncated protein rescues the GMC symmetric division phenotype in mam in an allele-specific manner. We further show that there is an interallelic complementation of loss-of-asymmetry defect between two alleles of mam. Our genetic results suggest that Mam protein might associate with itself during the asymmetric specification of cell fates and that the first two-thirds of the protein is essential in this process.
MATERIALS AND METHODS
Fly stocks and genetics: The following mam alleles were used in this study: ethyl methanesulfonate (EMS)-induced allele cn mamlL42bw sp/CyO, a spontaneous inversion mamN2G/CyO, and the hybrid dysgenic allele mamHD 10/6/CyO (Yedvobnicket al. 1988; Schmidet al. 1996). Heat shock-GAL4 on chromosome 3 was obtained from H. Keshishian. For the numb allele, we used nb796. nb796, mamIL42 double mutants were generated by recombination.
Heat shock-GAL4/UAS-Mam truncation strains: Construction of Mam truncations in pUAST is described in Helms et al. (1999). The Mam protein in the UAS-MamN transgene terminates at nucleotide 3884 of cDNA B4 (Smolleret al. 1990), encoding through the first acidic charge cluster, as well as an additional 500 residues, and ending at Mam residue 1043. UAS-MamH terminates at nucleotide 1489, Mam residue 245, which is 55 residues carboxy to the basic charge cluster. Germline transformants carrying either UAS-MamN or UAS-MamH on chromosome 3 were mated with a Hs-GAL4 chromosome 3 strain. Females trans-heterozygous for the transgenes were mated to w1118 and male recombinants (hs-MamN and hs-MamH) selected by eye color. Recombinant chromosomes were balanced over TM3 Sb and homozygous lines were selected. The strains were tested for phenotypic effects after heat-shock treatment (34° for 5 min) of third instar larvae. Both strains exhibited macrochaete duplications, eye defects, and loss of wing material (see results) consistent with dominant-negative effects of the Mam truncations described previously (Helmset al. 1999).
Heat-shock regimen during embryogenesis: Embryos were collected for 2 hr and then either heat-shocked immediately at 41° for 30 min or aged at various intervals and then heat-shocked (see text for details). Embryos were then aged again for different durations prior to fixation and then stained with anti-Eve or anti-Eve and anti-Zfh-1.
Rescue experiments: UAS-MamN and Hs-GAL4 transgenes were introduced into either the mamHD10/6 or the mamIL42 background. Embryos from these combinations were collected for 2 hr, aged 6 hr (6–8 hr old) and the UAS-MamN was induced by heat shock at 41° for 30 min. These embryos were allowed to grow at room temperature until they reached ∼14 hr or older before fixing and staining with anti-Eve.
Sequencing of the mam mutant alleles: mamIL42 homozygous embryos were identified by the presence of CNS defects (visualized with Eve staining) and lack of balancer-specific staining. DNA from several individual homozygous embryos were individually prepared and the mam coding region between MamH and MamN (see Figure 2a) was amplified from these individual DNA preparations. The amplified DNA was then sequenced in both directions.
Antibodies and immunostaining: Embryos were stained using standard immunohistochemistry procedures. Embryos were fixed and stained with Eve (1:2000), Eve and Zfh-1 (1:400), or LacZ (1:2000). For light microscopy, alkaline phosphatase or 3,3′-diaminobenzidine-conjugated secondary antibodies were used. For confocal, FITC and Cy5 secondary antibodies were used. Embryos of appropriate genotypes (i.e., mutants or rescue embryos) were identified using blue balancers and or marker phenotypes.
The GMC-1 → RP2/sib lineage: The GMC-1 → RP2/sib lineage, generated by NB4-2, is one of the well-studied neuronal lineages in the ventral nerve cord of the Drosophila embryo (reviewed in Bhat 1999). NB4-2 is delaminated in the second wave of NB delamination during midstage 9 (∼4.5 hr old) of embryogenesis (Hartenstein and Campos-Ortega 1984; Doe 1992) and is located in the fourth row along the anterior-posterior axis and in the second column along the medio-lateral axis within a hemi-segment. It generates its first GMC (GMC-1, also known as GMC4-2a) ∼1.5 hr after formation. The GMC-1 divides ∼1.5 hr later to generate two cells, the RP2 and the sib.
There are several well-established ways to distinguish a GMC-1, an RP2, and a sib (see Bhat and Schedl 1994; Buescheret al. 1998; Waiet al. 1999). First, the nuclear division and cytokinesis of GMC-1 that generates the two daughter cells is asymmetric in nearly 97% of the hemi-segments (the number of hemi-segments examined, N = 400). Thus, in 7.5- to 10-hr-old embryos, the cell that is destined to become an RP2 is significantly larger compared to the cell that will eventually become a sib (cf. Figure 1, b and c). Second is the level of marker gene expression between an RP2 and a sib as well as the temporal dynamics of expression of marker genes. For example, in nearly 99% of the hemi-segments, the future RP2 cell has a stronger expression of markers, such as Even-skipped (Eve), compared to the cell that is destined to become a sib (Figure 1, b and c). We have not encountered a newly formed sib that is the same size as a newly formed RP2 and has the same level of expression of marker genes as in an RP2 (N = >1000). Third, the cell that eventually assumes a sib identity undergoes a size reduction (Figure 1c) and further downregulation of expression of RP2-specific marker genes. By 13–14 hr of development, the sib loses Eve expression (cf. Figure 1d). Finally, RP2 is a motor neuron whereas sib has no axon projection and its eventual fate is unknown.
Loss of function for mam causes both loss-of-asymmetry and neurogenic phenotypes: Most mam alleles show a neurogenic phenotype (Yedvobnicket al. 1988; Schmidet al. 1996). We found that one of the EMS-induced alleles, mamIL42, showed a mild neurogenic phenotype whereas it had a strong loss-of-asymmetry phenotype. For example, the GMC-1 of the RP2/sib lineage divides symmetrically into two RP2s in as many as 80% of the hemi-segments, resulting in the duplication of RP2 neurons (Figure 1, f–h, arrows). A slight size asymmetry is present between the two RP2 neurons, which is also the case in Notch mutants (Waiet al. 1999). The smaller cell is also an RP2 in these mutants as indicated by its RP2-specific axon projection pattern (data not shown; cf. Waiet al. 1999). The other alleles examined all had a strong neurogenic phenotype (cf. Figure 2f); however, these alleles also had the loss-of-asymmetry phenotype, exemplified by the symmetric division of GMC-1 into two RP2 neurons (Figure 2f, arrows).
Expression of the MamH truncation interferes with both the neurogenic and the asymmetry functions of wild-type Mam: Mam protein has several distinct domains, such as an N-terminal basic domain, a centrally located acidic domain, and a C-terminal acidic domain (Figure 2a). First, using a transgenic line carrying an N-terminal truncation of Mam (MamH; see Figure 2a) under the control of upstream activator sequence (UAS; Helmset al. 1999), we ectopically expressed this truncated Mam using GAL4 under the control of the heat-inducible heat shock 70 gene promoter (Hs-GAL4) at different developmental time points. Early induction of this transgene (between 2 and 6 hr of development at 22°) resulted in a neurogenic phenotype (Figure 2b). When the transgene was induced during the time in which the GMC-1 of the RP2/sib lineage undergoes asymmetric division (between 6 and 8 hr of development at 22°), it appeared that GMC-1 divided symmetrically into two RP2 neurons (Figure 2c). However, with the MamH transgenic line, we encountered a problem: following induction, the embryos failed to retract germ band and therefore we could not ascertain if both the daughters of GMC-1 adopt an RP2 fate by staining for Eve expression alone. Therefore, we double stained these embryos with Eve and Zfh-1. Zfh-1 is a zinc-finger protein and is expressed in a newly formed RP2 but not in a GMC-1 or a sib (Waiet al. 1999; Mehta and Bhat 2001). Double staining of embryos with Eve and Zfh-1 where the MamH transgene was induced between 6 and 8 hr of development revealed that both the progeny of GMC-1 have Eve and Zfh-1 (Figure 2e, arrows). The frequency of loss of asymmetry was low (∼10% of the hemisegments). Nonetheless, these results indicate that this truncated form of Mam functions as a dominant negative competing with the wild-type Mam and produces both neurogenic and loss-of-asymmetry phenotypes. We also examined an allele of mam, N2G, in which an inversion breaks the gene in such a way that it is expected to produce a truncated form of the protein that is 12 amino acids shorter than MamH (Figure 2a). In this allele, we observed both the neurogenic and the loss-of-asymmetry phenotypes (Figure 2f).
Expression of the MamN truncation interferes with the neurogenic function of wild-type Mam but not with its asymmetry function: We next examined a transgenic line carrying a longer form of the Mam protein, MamN (see Figure 2a; Helmset al. 1999). When this transgene was induced using the Hs-GAL4 driver during early neurogenesis (between 2 and 6 hr of development at 22°), it resulted in a neurogenic phenotype with a large number of neurons forming at the expense of ectoderm (Figure 2, g and h). A similarly truncated version of human Mam produces a neurogenic effect in Xenopus (Fryeret al. 2002). However, when the MamN transgene was induced between 6 and 8 hr of development at 22°, unlike the MamH, it had no effect on GMC-1 division (Figure 2i).
Loss of Mam or Notch activity in imaginal discs has been shown to cause adult phenotypes (see Helmset al. 1999 and references therein). These include macrochaete duplications, eye defects, and wing defects. Induction of Mam truncations in imaginal discs has been shown to cause similar phenotypes (Helmset al. 1999). We repeated these experiments to make sure that the newly constructed Hs-GAL4, MamH, and Hs-GAL4, MamN recombinant chromosomes behave similarly. As shown in Figure 3, a brief induction of the MamH transgene during the third instar larval stage produced macrochaete duplications, eye scarring, and wing defects. Similar results were observed with the induction of the Hs-GAL4, MamN transgene (data not shown). These results further indicate that these transgenes behave as loss-of-function mam mutations.
MamN rescues the loss-of-asymmetry phenotype in a hypomorphic mam allele: The above results indicate that MamN can interfere with wild-type Mam function during neuroblast formation and thus induce a neurogenic phenotype, whereas it cannot interfere with the wild-type protein during asymmetric fate specification. We hypothesized that perhaps MamN has the part of the protein required for specifying sib identity and thus it does not function as a dominant negative during asymmetric fate specification. To test this hypothesis, we sought to rescue the loss-of-asymmetry phenotype in two different alleles, mamHD10/6 and mamIL42. mamHD10/6 is a P-element insertion allele in which the P element is inserted in the untranslated first exon (Smolleret al. 1990). This allele shows the loss-of-asymmetry phenotype (Figure 4B; see also Table 1) but not the neurogenic phenotype (Waiet al. 1999); mamHD10/6 does show a neurogenic phenotype in combination with stronger alleles (Yedvobnicket al. 1988). The Mam protein in this allele is predicted to be wild type but most likely present at reduced levels. mamIL42 is an EMS-induced allele and this produced a mild neurogenic phenotype but a strong loss-of-asymmetry phenotype (Figure 4D; see also Figure 1 and Table 1). We introduced MamN into these mutant backgrounds and induced the gene using Hs-GAL4 between 6 and 8 hr of development. When these embryos were examined for the RP2/sib lineage division pattern, we found that the asymmetry division defect in the RP2/sib lineage in mamHD10/6 was rescued by MamN (Figure 4C). However, the asymmetric division defect (as well as the neurogenic phenotype) in mamIL42 was not rescued by MamN (Figure 4E). These results suggest that MamN contains the regions necessary for generating asymmetry but in an allele-specific manner. Given this allele-specific difference in the ability of MamN to rescue the defects, we sequenced the portion of the mam gene implicated in this function in the mamIL42 allele. As shown in Figure 5a, this allele had a change from glutamine at position 1038 (within a stretch of glutamines) to a stop codon, predicting a truncated protein seven amino acids shorter than MamN.
mamIL42 and mamHD10/6 show interallelic complementation: As discussed above, ectopic expression of MamN elicits a neurogenic defect but not the loss-of-asymmetry phenotype. Further, MamN can rescue the loss of asymmetric division in mamHD10/6 but not in mamIL42. These results led us to think that MamN can rescue the asymmetric division defect only in the presence of some wild-type Mam protein. This raised the possibility that MamN might interact with wild-type Mam protein during the specification of sib fate. We tested this idea genetically by looking for interallelic complementation. When two molecules of the same protein interact with one another, this is often revealed by interallelic complementation (cf. Tanget al. 1998). Therefore, we crossed mamHD10/6 to mamIL42 and examined the trans-heterozygous embryos. As shown in Figure 4F, there was a significant rescue of the asymmetric division defect and thus most hemi-segments had normal specification of sib (see Table 1). This interallelic combination is similar to the mamHD10/6; MamN combination and the rescue of asymmetric division defect in these combinations indicates that MamN or MamIL42 associates with Mam to rescue the loss-of-asymmetry defect.
mamIL42 carries a suppressor of the neurogenic defect but not the asymmetry defect: The strong neurogenic effect of MamN but the absence of a similar strong neurogenic defect in mamIL42 (which is predicted to produce a truncated MamN-like protein; see Figure 5a) was unexpected. For example, the absence of a strong neurogenic defect in mamIL42 indicates that the truncated protein has all the necessary function for normal NB formation. However, the dominant-negative neurogenic effect of MamN shows that this truncated protein does not carry the function necessary for normal NB formation. While MamN carries an additional seven amino acids compared to MamIL42, we considered the possibility that the mamIL42 mutant carries a suppressor(s) of the neurogenic defect but not a suppressor(s) of loss-of-asymmetry defect. Since a straightforward outcrossing of mamIL42 does not result in the loss of suppressor(s), the suppressor(s) are likely to be located on the same chromosome as mamIL42. Therefore, we subjected mamIL42 to one round of recombination and examined embryo collections from recombinants. Consistent with the possibility of the presence of a partial suppressor of neurogenic defect, we recovered mamIL42 chromosomes that showed a strong neurogenic defect (Figure 5c). These embryos, however, showed the loss-of-asymmetry defect to the same extent as the embryos from the original mamIL42 strain (Figure 5c; see Table 1). We have not yet mapped the suppressor(s) nor have we determined that the effect is due to a single locus.
mam phenotype is epistatic to the numb phenotype: Recent studies indicate that inscuteable (insc) and nb play a crucial role in the terminal asymmetric division of GMC-1 of the RP2/sib lineage (Buescheret al. 1998; Learet al. 1999; Waiet al. 1999). The asymmetric divisions mediated by these proteins appear to be tied to their asymmetric localization in GMC-1 and to their asymmetric segregation between two daughter cells during division. For instance, during the division of GMC-1 of the RP2/sib lineage, Insc localizes to the apical end of GMC-1, which in turn segregates Nb to the basal end. The cell that inherits Nb is specified as RP2 due to the ability of Nb to block Notch signaling, whereas the cell that does not inherit Nb (but inherits Insc) is specified as sib by Notch. Thus, in insc mutants, both daughters of the GMC-1 adopt an RP2 fate whereas in nb mutants they assume a sib fate (Buescheret al. 1998; Waiet al. 1999). Our previous results indicate that the sib cell adopts an RP2 fate in Notch; nb double mutants (Waiet al. 1999), indicating that Nb is needed to specify RP2 fate only when there is intact Notch. We sought to determine if the same relationship exists between mam and nb. We generated mam, nb double mutants and examined the division pattern of GMC-1. For this purpose, we used the allele mamIL42. This mam allele, although not a null, shows a very strong loss-of-asymmetry phenotype (83% of the hemi-segments; see Table 1), which can be reliably identified. A null allele gives a severe neurogenic phenotype and makes it far more difficult to determine the double-mutant phenotype. In addition, we had determined the molecular lesion in the mam gene in this allele. Since this allele is a loss-of-function allele and the phenotypes of mam and numb mutants are opposing phenotypes, we reasoned that use of non-null alleles should not pose any problems for analysis or interpretation of the double-mutant results. Therefore, we used this allele in our double-mutant experiments. As shown in Figure 6c, both the daughter cells of GMC-1 adopted RP2 fate in these embryos.
The canonical Notch pathway described for Drosophila (Artavanis-Tsakonaset al. 1999) is widely conserved, including Caenorhabditis elegans and mammals. The intercellular communication mediated by Notch involves interactions between Delta, the signal, and Notch, the receptor (Fehonet al. 1990; Heitzler and Simpson 1993). The signaling involves transendocytosis of the Notch extracellular domain bound to Delta into the signaling cell (Parkset al. 2000). A physical perturbation of Notch protein structure during transendocytosis may be needed for proteolytic processing and release of the Notch intracellular domain (Notchintra; Mumm and Kopan 2000). Proteolytic cleavage of Notch is mediated by the membrane-bound Presenilin protein (Yeet al. 1999). In many but not all contexts (Ramainet al. 2001), signaling by Notch occurs in conjunction with the DNA-binding Su(H) protein, the mammalian CBF1/worm Lag-1 homolog (CSL; Henkelet al. 1994). In the absence of Notch signaling, Su(H) establishes a default state of gene repression, which appears to be mediated via complexing with Hairless and the corepressors Groucho and dCtBP (Baroloet al. 2002). Upon Notch activation, Nintra goes into the nucleus, where it dissociates the repression complex and leads to the formation of an activation complex containing Su(H) and Mam.
In the ventral nerve cord of the Drosophila embryo, the Notch pathway mediates terminal asymmetric division of secondary neuronal precursor cells (Buescheret al. 1998; Learet al. 1999; Schuldt and Brand 1999; Waiet al. 1999). The secondary precursor cells, GMCs, in the nerve cord generally divide by asymmetric mitosis to generate two different daughter cells. For example, in the GMC-1 → RP2/sib lineage during GMC-1 division, the Inscuteable protein asymmetrically localizes to the apical end, which forces Numb to localize to the basal end. Basally localized Numb then segregates to the future RP2. The function of Numb is to prevent the cleaving of the intracellular domain of Notch. In the absence of Numb, the intracellular domain of Notch gets cleaved and then translocated into the nucleus where it specifies a sib fate by complexing with Su(H) and Mam and activating downstream target genes. Previous results also show that for the specification of an RP2 identity Numb is not required, but it is required to prevent that cell from becoming a sib in the presence of an intact Notch pathway.
In this article, we show differential effects of Mam on asymmetric cell fate specification vs. neuroblast formation in the ventral nerve cord of the Drosophila embryo. We show that a Mam truncation, which has the basic and the first acidic domain (MamN), rescues the asymmetric cell fate specification defect in an allele-specific manner. These conclusions are based on several lines of evidence. First, a transgene that encodes this truncated Mam protein causes a dominant-negative neurogenic defect, but it does not cause a dominant-negative effect on asymmetric division. Thus, expression of this transgene during the asymmetric division of GMC-1 does not cause a duplication of RP2 as one would expect if this transgene functions as a dominant negative. The same transgene when expressed earlier when NBs are formed causes a neurogenic defect. This indicates that the truncated transgene functions as a dominant negative but only during the earlier neurogenic process. Second, MamN rescues the asymmetry defect in one of the mam mutant alleles, mamHD10/6. This is a hypomorphic P-element insertion allele (Smolleret al. 1990), which causes the loss of asymmetric division defect (Waiet al. 1999) but does not cause a neurogenic defect except in combination with strong alleles of mam (Yedvobnicket al. 1988). These results and the fact that the P element is inserted in the untranslated first exon suggest that low levels of wild-type Mam are produced by this allele. However, the finding that MamN does not rescue the asymmetry defect in another mam mutant allele, mamIL42, which is predicted to produce a truncated Mam protein similar to MamN, indicates that this rescue is allele specific (see below). Thus, some wild-type Mam protein appears to be necessary for the rescue by MamN and it is possible that the two proteins interact to provide the rescue function (see below).
Our sequence analysis of mamIL42 suggests that this allele encodes a Mam protein that is similar to MamN (although it is seven amino acids shorter). The inability of MamN to rescue mamIL42 argues that this truncated protein in combination with MamN is not sufficient to rescue the asymmetry defect. However, the interallelic complementation between mamHD10/6 and mamIL42 (a situation very similar to the mamHD10/6;MamN combination) also suggests that MamIL42 and MamHD10/6 proteins (which are expected to be wild type, but present at reduced levels) interact to rescue the loss of asymmetric division of GMC-1. These results raise the question as to whether or not MamN (which is similar to the Mam protein in the mamIL42 allele) has all the necessary function for generating asymmetry. Since it does not rescue the asymmetry defect in mamIL42, clearly it does not have all the necessary information. However, it does have the required function in the presence of some presumably wild-type protein (i.e.,in mamHD10/6 background). This is consistent with the fact that MamN does not function as a dominant negative during the asymmetric division of GMC-1 but only at earlier stages during the formation of NBs.
There might be some difference between MamN and MamIL42 in their ability to complement loss of asymmetric division in mamHD10/6. This is indicated by the findings that while MamN can rescue the asymmetry defect in mamHD10/6, the interallelic complementation of the asymmetry phenotype between mamIL42 and mamHD10/6 is not as complete as rescue of mamHD10/6 by MamN (see Table 1). This may, in part, be due to the seven-amino-acid difference between MamN and MamIL42. Alternatively, there may be a protein-level difference between the two cases; in the former, MamN is expressed at high levels under Hs-GAL4, whereas in the latter mamIL42 is under the control of the mam promoter. Yet, the seven-amino-acid residues could make some difference, given that these amino acids are mostly glutamine residues, which can be involved in multimerization of proteins (Pascal and Tjian 1991). It is possible that the region of the Mam polypeptide defined by MamN (and MamIL42) is required to interact efficiently with the full-length Mam during the asymmetric fate specification. The requirement of some wild-type Mam protein for the rescue activity of MamN or MamIL42 also suggests that the remaining portions of Mam are also required for generating asymmetry. The most likely scenario would be that this is a protein-protein interaction, although some other possibilities cannot be excluded. Since the available antibody against Mam recognizes multiple bands on a Western blot of proteins from embryo, we have not performed immunoprecipitation experiments to address protein-protein interaction between Mam molecules.
Our results show that mamIL42 carries a partial suppressor of neurogenic defect since a strong neurogenic defect can be restored to this allele upon recombination. This is consistent with the result that expression of MamN elicits a strong dominant-negative neurogenic defect. However, this suppressor in mamIL42 has no modifying effect on the loss-of-asymmetry phenotype of mamIL42, as indicated by the fact that there was no change in the penetrance of this defect between the original and the recombinant mamIL42. We have not mapped the location of this suppressor(s) beyond its tentative assignment to chromosome 2.
Previous studies utilizing MamH and MamN have demonstrated that both truncations elicited dominant-negative effects when overexpressed in imaginal tissues (Helmset al. 1999). It was later shown that the basic region of Mam is conserved in fly, mouse, and human and that the region physically interacts with the processed intracellular segment of Notch (Nintra; Petcherski and Kimble 2000a,b; Wuet al. 2000; Kitagawaet al. 2001). Mam, Nintra, and Su(H)/CSL proteins associate in a ternary complex that binds to HES/E(spl) promoters and activates gene expression (Wuet al. 2000; Kitagawaet al. 2001). The expression of MamH and MamN presumably leads to transcription complexes containing a defective form of Mam in a complex with Su(H) and Nintra. The current results, however, indicate that these interactions may be distinct during the generation of asymmetry. For instance, the MamN polypeptide may lack sequences required for interaction with factors necessary for NB formation but not for asymmetric division.
Finally, our results indicate that the mam phenotype in the RP2/sib lineage (symmetrical division of GMC-1 into RP2 and sib) is epistatic to the numb phenotype (symmetrical division of GMC-1 into two sibs). During the division of GMC-1, Insc localizes to the apical end of GMC-1, which in turn segregates Nb to the basal end. The cell that inherits Nb is specified as RP2 due to the ability of Nb to block Notch signaling, whereas the cell that does not inherit Nb (but inherits Insc) is specified as sib by Notch. Thus, in insc mutants, both daughters of the GMC-1 adopt an RP2 fate whereas in nb mutants they assume a sib fate (Buescheret al. 1998; Waiet al. 1999). The sib cell adopts an RP2 fate in Notch; nb double mutants (Waiet al. 1999). This indicates that Nb is needed to specify RP2 fate only when there is intact Notch. The mam, numb double mutant result is consistent with the above result and extends our previous finding. That is, Numb is needed only when there is intact Mam. This result further indicates that Mam functions downstream of Notch during the asymmetric specification of RP2 and sib, an observation consistent with the prevailing view of the Notch signal transduction pathway.
We thank Manfred Frasch and Zhi-Chun Lai for the generous supply of antibodies and Spyros Artavanis-Tsakonas, Haig Keshishian, and the Bloomington Stock Center for various stocks. Comments from members of the Bhat lab and Yedvobnick lab were appreciated. B.Y. wishes to thank K.B. and the members of the Bhat lab for generously sharing their time and expertise during his time in their lab. This work is supported by a grant from the National Science Foundation to B.Y. (IBN 9904411) and a grant from the National Institutes of Health to K.B. (GM58237).
Communicating editor: R. S. Hawley
- Received July 7, 2003.
- Accepted November 24, 2003.
- Copyright © 2004 by the Genetics Society of America