Notch Group of Neurogenic Genes

Phenotypic and genetic analyses demonstrate that fs( 1) Yb activity is required in the soma for the development of a subset of ovarian follicle cells and to support later stages of egg maturation. Mutations in f s ( 1 ) Yb cause a range of ovarian phenotypes, from the improper segregation of egg chambers to abnormal dorsal appendage formation. The mutant phenotypes associated with fs( I) Yb are very similar to the ovarian aberrations produced by temperature-sensitive alleles of Notch and Delta. Possible functional or regulatory interactions between fs(1) Yb and Notch are suggested by genetic studies. A duplication of the Notch locus partially suppresses the female-sterility caused byfs(1) Ybmutations, while reducing Notch dosage makes the fs(1) Yb mutant phenotype more severe. In addition, fs( 1) Yb alleles also interact with genes that are known to act with or regulate Notch activity, including Delta, daughterless, and mastermind. However, differences between the mutant ovarian phenotype of fs( 1) Yb and that of Notch or Delta indicate that the genes do not have completely overlapping functions in the ovary. We propose that fs(1) Yb acts as an ovary-specific factor that determines follicle cell fate. 0 OGENESIS in Drosophila requires that both germline and somatic cells differentiate in a coordinated manner (KING 1970; SPRADLING 1993). The Drosophila ovary is subdivided into ovarioles, each of which contains a row of progressively developing egg chambers (Figure 1A). New egg chambers, made up of 16 germ cells surrounded by a layer of somatic follicle cells, are continuously produced in an apical structure called the germarium. The egg chambers grow and develop as they move through the ovariole, reaching the oviduct as a mature egg. The follicle cells have specialized fates that are essential for the organization of the egg chamber and the continued development of the germ cells. Follicle cells on the anterior side of the cyst delaminate and act to pinch off the new egg chamber from the germarium (Figure 1B) . These follicle cells go on to form a structure called the stalk that acts to connect and separate the egg chambers as they move through the ovariole. The delaminating follicle cells also give rise to the specialized polar cells that were identified by their distinct morphology and the localized expression of certain enhancer trap lines (KING 1970; GROSSNIKLAUS et al. 1989). Two polar cells are found at each end of the egg chamber and may be required for the reorganization of microtubules along the anterior-posterior axis, suggestive of a direct role for these cells in establishing polarity within the egg chamber (RUOHOLA et al. 1991; CLARK Curresponding author: Rod Nagoshi, Department of Biological Sciences, The University of Iowa, Iowa City, IA 52242. Genetics 140: 207-217 (May, 1995) et al . 1994). The anterior polar cells are a subset of a larger group of follicle cells called border cells. These migrate from the anterior end of the egg chamber to the oocyte-nurse cell border and have a role in forming the micropyle ( MONTELL et al. 1992) . Additional follicle cell movements include the posterior migration within stage 9 egg chambers and the migration of the centripedal follicle cells to cover the anterior end of the egg chamber ( SPRADLING 1993). A group of genes that were first identified as controlling embryonic neurogenesis ( LEHMANN et al. 1983), also have a role in specifjmg cell fate in the follicle layer ( RUOHOLA et al. 1991 ) . The earliest identified member of this group, Notch (N), is a large transmembrane protein with epidermal growth factor (EGF) repeats in its extracellular domain ( WHARTON et al. 1985; KIDD et al. 1986). Other loci with similar embryonic phenotypes have also been identified and, along with N, are known as the neurogenic genes. One of these genes, Delta ( D l ) , also encodes a transmembrane protein with EGF repeats (VASSIN et al. 1987; KOPCZYNSKI et al. 1988) , and has been shown in cell aggregation assays to physically interact with the Nproduct ( FEHON et al. 1990). Nand D l also interact genetically, as increasing the dosage of Dl enhances the severity of the loss-of-function N embryonic phenotype (DE LA CONCHA et al. 1988) , Similar genetic interactions are seen between many of the neurogenic genes, suggesting that they function together in a common process. The mechanism by which the neurogenic genes con208 E. Johnson, S. M'avnc and R. Nagoshi FIGURE 1.-Normal dcvclopmcntal stages o f ' ooynrsis. Single ovarioles wcrc disscctrtl f rom an o\.ary and staincd with the nuclear dye DAH. ( A ) The most apical structure, thc germarium, is to the left. A mature egg chamber is to the right. (R) A stage 2 egg chamber pinching off from the germarium. Figure 1R is at a fourfold greater magnification than Figure 1A. g, germarium; f, follicle cells; n, nurse cells; 0, oocyte; 2, stage 2 egg chamber; p, region of pinching off; s, stalk. trol cell fate choice is thought to involve cellcell interactions. Embryos lacking N or Dl function have an increased number of ectodermal cells that take up a neural fate, instead of their normal epidermal cell fate (for review see CAMPOS-ORTEGA 1993). The process of segregating neuroblasts from a common pool of precursors is thought to be mediated by lateral inhibition, a process by which a cell committed to a specific develop mental fate inhibits surrounding cells from becoming similarly determined (DOE and GOODMAN 1985; ARTAVANIS-TSAKONAS and SIMPSON 1991). Based on this model, Nand Dl mutant ectodermal cells can no longer receive or send the signal that inhibits neural determination, thereby leading to an increased number of neural cells. Mosaic analyses demonstrate that Dl is nonautonomous in its function while N is cell-autonomous, consistent with N acting as a receptor for an intercellular inhibitory signal encoded by Dl ( HEITZLER and SIMPIn addition to their role in neurogenesis, many of the neurogenic genes are required for egg development. Female flies that are mutant for N, Dl, or dnugh!prlhs (da) in the follicle cells, or irniniac (im) in the germline, are sterile (RUOHOLA et nl. 1991; Goom et nl. 1992; XU et al. 1992; BENDER et 0.1. 1993; CUMMINGS and CRONMILLER 1994). These genes are needed for proper envelopment of the germline cysts by the follicle cells, apparently by regulating follicle cell specification. N and Dl mutant ovaries have a hyperplasia of polar cells at the expense of stalk cells. By analogy with their lateral inhibitory role in neurogenesis, Nand Dl may act in a subset of the precursor follicle cell population to inhibit polar cell determination, thereby allowing the formation of stalk cells. The Notch group of regulatory genes not only control cell fate decisions in neurogenesis and oogenesis but also have similar functions in tissues as diverse as adult retina ( CACAN and READY 1989), muscle ( CORRIN et SON 1991 ). nl. 1991 ) , and sensory bristles ( GHYSEN and DAMRLYCHAUDIERE 1989). This indicates that the same basic framework of gene products, a "gene cassette", can be used to control different cell fate decisions ( RERAY et nl. 1991; ART.4VANIS-TSAKONAS and SIMPSON 1991; RUOHOIA-BAKER pt nl. 1994). In this case, a gene cassette made up of members of the Notch group might be generically used to inhibit the competence of cells to commit to a particular developmental fate (FORTINI and ARTAVANIS-TSAKONAS 1993), be it to form polar cells among follicle precursors or to become neural in an ectodermal cell population. For a gene cassette to function in multiple developmental pathways, tissue-specific factors are required to interpret the generic regulatory signals from the cassette in a manner that makes sense for that particular cell type. In this paper, we characterize a soma-specific function required for ovarian development. The Ji( I ) Yh gene was initially described by mutations that caused semisterility in females (YOUNG and JUDD 1978). We have identified additional, more severe alleles, that show substantial dosage-sensitive interactions with Notch. We demonstrate that the Yh and Notch functions are required for the development of overlapping subsets of follicle cells. MATERIALS AND METHODS Fly stocks: Flies were raised at 29" on standard cornmeal, molasses, yeast, agar media containing propionic acid as a mold inhibitor and supplemented with live yeast. The new Js( 1 ) YI, alleles were originally isolated on a y cu v f parental chromosome and called Js( 1) M104 ( MOHIXR and CARROI.I. 1984). We replaced most of the proximal portion of the X chromosome with nonmutagenized sequences to create y J s ( l ) Y h 4 f and y J$( l )YI , 'J: An existing J s ( l ) Y b allele, y z Ji(1)YI)' &spI sn3 (YOUNG and Junn 1978), and the 8-kb genomic insertion stock pW5Jd. l are described in LIN and M'Ol.FNER (1989). Two other J s ( 1 ) YI, alleles have been described previously (YOUNG andJunn 1978), but appear to be fs( 1 ) Yb Action During Oogenesis 209 lost. The N + duplication stock ( y w f ; Dp( 1:2) 51b) and a loss-of-function daughtmhss ( d a ) allele (stock: da' pcn/SM5) were obtained from the Bloomington Stock Center. Loss-offunction alleles of Delta ( D l , stock: Dlx/Payne) , mastermind (mam, stock: cn mum'B99 bw sp /CyO) , big brain (bib, stock: Zn( 1 ) dl-49, y Hw mz were obtained from the Mid-America Drosophila Stock Center. C ( l ) A , y / + ; cosP479LE ( N + ) / TM6B is a transgenic cosmid construct containing the intact Notch gene (gift from SPYROS ARTAVANIS-TSAKONAS) . Two enhancer trap lines were used in this study: c368 is an enhancer trap line specific for stalk cells and was a gift of L. MANSEAU. Enhancer trap line pAlOl.F3/ CyOis specific for the neuralized gene ( neur) and was a gift of H. BELLEN. All deletions and lethals used in the mapping study were obtained from the Bloomington Stock Center. Viability and fecundity crosses: For the viability tests, y f s ( 1 ) Yb" cv u f/FM7females were mated to wildtype ( OregonR ) males and y f s ( 1 ) Yb" cu v f / + daughters collected. These were then crossed to y fs( 1 ) Yb6 cv v f males. In the ro en of this cross, we compared the number of y fs( 1 ) Yb cu u f / y f s ( l ) Y b 6 c v v fdaughterstotheiryfs(l)Yb6cvvf/+sisters and the number of y fs ( 1 ) Y b 4 cu v f / Y sons compared with their wild-type brothers. The same crosses were made using the y cv u f stock from which fs( 1 ) Yb" was isolated to control for background effects. Additional crosses using the y fs ( 1 ) Yb4 f and y fs( 1 ) Yb6 f chromosomes followed the same scheme. To determine the effect of the mutations on male fertility and fecundity, we compared the number of progeny produced from&( 1 ) Yb" males with wild-type males. Fifty$( 1 ) Yb" males were each mated singly to three wild-type females. The same number of crosses were made with y cu v f males. The frequency of fertile males and the number of progeny produced by each male were determined for the two genotypes. Germline clonal analysis: The germline dependence of f s ( 1 ) Yb was tested by examining f s ( 1 ) Yb mutant germline clones produced by the dominant female-sterile technique ( PERRIMON and GANS 1983). The dominant mutation ovoDl blocks oogenesis if present in one copy in the germline. To induce mitotic crossing over, y f s ( 1 ) Y b ouo+ cu v f / f s ( l ) Y b + ouol)' u flies were irradiated with 1000 R of gamma radiation from a 137Cs source during the early larval period, which is expected to produce clones occupying several ovarioles ( WIESCHAUS and SZABAD 1979). Females were mated to y cv v f / Y males and tested for fertility. Confirmation that the fertile germline clones were fs ( 1 ) Ybcame by examining the progeny for expression of the y and cv markers. Ovary staining: Ovaries were stained with the nuclear dye DAPI for the examination of ovary morphology. Ovaries were dissected from 3to 5day adult females in phosphate-buffered saline (PBS; 150 mM NaCI, 14 mM Na2HP04, 6 mM NaHpP04, pH 7.0) and then fixed in a 1:l mixture of 4% paraformaldehyde in PBS: heptane for 5 min with constant shaking. After rinsing with PBS, ovaries were stained with 0.5 pg/ml DAPI in PBS for 30 min. The ovaries were then washed three times with PBS and mounted in 50% glycerol in PBS. Stained ovaries were illuminated by epifluorescence with a UV filter. Enhancer trap expression patterns were visualized by staining for 0-galactosidase activity. Ovaries were fixed as above, then stained in PBS + 5 mM KsFe ( CN ) + 5 mM w e ( CN ) + 0.2% 5-bromo4chloro-3-indoxyl-~-galactopyranoside (Xgal). Ovaries were then rinsed in PBS and mounted as above. RESULTS hblD05 cn bw sp/CyO) , and Notch ( N , stock f a Ni2"' sn3/