Identification of X-Linked Genes Required for Migration and Programmed Cell Death of Drosophila melanogaster Germ Cells

Corresponding author: Clark R. Coffman, 3258 Molecular Biology Bldg., Iowa State University, Ames, IA 50011-3260., ccoffman{at}iastate.edu (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila germ cells form at the posterior pole of the embryo and migrate to the somatic gonad. Approximately 50% of the germ cells that form reach their target. The errant cells within the embryo undergo developmentally regulated cell death. Prior studies have identified some autosomal genes that regulate germ cell migration, but the genes that control germ cell death are not known. To identify X-linked genes required for germ cell migration and/or death, we performed a screen for mutations that disrupt these processes. Here we report the identification of scattershot and outsiders, two genes that regulate the programmed death of germ cells. The scattershot gene is defined by a mutation that disrupts both germ cell migration and the death of germ cells ectopic to the gonad. Maternal and zygotic expression of scattershot is required, but the migration and cell death functions can be genetically uncoupled. Zygotic expression of wild-type scattershot rescues germ cell pathfinding, but does not restore the programmed death of errant cells. The outsiders gene is required zygotically. In outsiders mutant embryos, the appropriate number of germ cells is incorporated into the gonad, but germ cells ectopic to the gonad persist.

CELL migration and programmed cell death play critical roles in animal development, immune system function, wound healing, angiogenesis, and metastasis (LAUFFENBURGER and HORWITZ 1996 ; JACOBSON et al. 1997 ; HOLDER and KLEIN 1999 ; MEIER et al. 2000 ). Many cells are formed in one part of the body and must migrate to their ultimate locations to function. A common strategy is to overproduce cells and then eliminate those that are no longer needed or that are potentially dangerous to the animal. The control of cell movements and cell death in an organism must be precisely regulated, and these processes require the integration of a wide range of signals between and within cells.

The germ cells of Drosophila provide an excellent system for the study of cell migration and cell death. The movements of the germ cells have been well documented, and the elimination of ectopic and/or supernumerary germ cells occurs with great efficiency (SONNENBLICK 1941 , SONNENBLICK 1950 ; COUNCE 1963 ; FULLILOVE and JACOBSON 1978 ; UNDERWOOD et al. 1980 ; TECHNAU and CAMPOS-ORTEGA 1986 ; HAY et al. 1988 ; LASKO and ASHBURNER 1990 ; SMITH et al. 1992 ; JAGLARZ and HOWARD 1994 , JAGLARZ and HOWARD 1995 ; WARRIOR 1994 ; WILLIAMSON and LEHMANN 1996 ; CAMPOS-ORTEGA and HARTENSTEIN 1997 ; MOORE et al. 1998 ; WYLIE 1999 , WYLIE 2000 ; STARZ-GAIANO and LEHMANN 2001 ). Selected features of germ cell movements and the elimination of germ cells ectopic to the gonad are shown in Fig 1. Prior studies have shown that 50% of the primordial germ cells that form successfully complete migration and are incorporated into the somatic gonad (SONNENBLICK 1950 ; UNDERWOOD et al. 1980 ; TECHNAU and CAMPOS-ORTEGA 1986 ). Primordial germ cells that have been labeled using horseradish peroxidase or radioactive thymidine do not transdifferentiate, and very few transplanted cells persist outside of the gonads (UNDERWOOD et al. 1980 ; TECHNAU and CAMPOS-ORTEGA 1986 ). Collectively, these studies demonstrate that germ cells ectopic to the gonad die and that a mechanism regulating germ cell survival must exist within the embryo.

View larger version (62K): In this window In a new window Download PPT slide   Figure 1. Dynamics of germ cell development. Some of the stages of germ cell development are shown. The germ cells are labeled utilizing a fat facets-lacZ reporter transgene. Anterior is to the left in all panels. (A, B, and D) Lateral views. (C, E, F, and G) Dorsal views. (A) Germ cells form at the posterior pole of the embryo. (B and C) Germ band extension is complete and the germ cells are contained within the posterior midgut epithelium. (D and E) Germ cell migration through the posterior midgut epithelium is almost complete, and the germ cells are separating laterally into two populations. (F) Germ cells have migrated to and are coalescing with the somatic gonad precursor cells in a germ band retracted embryo. (G) Germ cell migration is complete. Note the absence of labeled cells outside of the gonads.

To date, no X-linked genes with roles in germ cell migration or programmed cell death have been reported. Screens of the second and third chromosomes identified some genes necessary for germ cell migration in Drosophila, but their roles remain enigmatic (KOBAYASHI et al. 1996 ; ZHANG et al. 1996 ; BOYLE et al. 1997 ; ZHANG et al. 1997 ; ASAOKA et al. 1998 ; BROIHIER et al. 1998 ; FORBES and LEHMANN 1998 ; MOORE et al. 1998 ; VAN DOREN et al. 1998 ; DESHPANDE et al. 1999 , DESHPANDE et al. 2001 ; STARZ-GAIANO et al. 2001 ). Our understanding of the regulation of germ cell development is far from complete, as this process must involve the functions of additional genes. Therefore, we conducted a screen of the X chromosome for genes that are required for normal germ cell development. When mutated, all of the genes identified in our screen altered the distribution of germ cells in the early Drosophila embryo. These mutations are informative to the study of germ cell migration and/or cell death, as they do not map to loci with previously described roles in these processes.

We report the identification and initial characterization of two genes. The first, scattershot (sctt), is maternally and zygotically required. This gene, when mutated, severely disrupts both germ cell migration and developmental cell death. In embryos from sctt/sctt mutant mothers, very few of the germ cells successfully migrate to the gonad. In addition, those germ cells ectopic to the gonad retain germ cell characteristics. Although the germ cell migration in sctt mutants is severely altered, germ cell formation, somatic gonad development, and body pattern appear normal. Homozygous female and hemizygous male sctt mutant embryos are viable and fertile, although fertility is greatly reduced. The sctt gene is only the third gene identified in Drosophila with a demonstrated role in germ cell migration whose product is maternally contributed. The other two are polar granule component (pgc) and nanos (nos; KOBAYASHI et al. 1996 ; NAKAMURA et al. 1996 ; ASAOKA et al. 1998 ; FORBES and LEHMANN 1998 ; DESHPANDE et al. 1999 ). Interestingly, zygotic expression of a wild-type copy of sctt rescues the germ cell migration defect, but fails to rescue the programmed cell death phenotype of sctt mutant embryos. A second gene, outsiders (out), must be expressed zygotically and has a function in the death of germ cells that fail to reach the gonad. The number of germ cells that reach the gonad is the same in out/out embryos and in unmutagenized controls. However, many additional cells outside of the gonad continue to express germ cell markers. Thus, both sctt and out are necessary for the execution of a cell death program, and sctt has an additional role in germ cell pathfinding.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutagenesis and screening:
To isolate X-linked mutations affecting germ cell development, 3475 independent mutagenized lines were screened. A P[w⁺, fat facets-lacZ] transgene that is expressed specifically in the germ cells of the embryo was used to determine the location of the germ cells (FISCHER-VIZE et al. 1992 ). The screen is diagrammed in Fig 2. Three- to 4-day-old w¹¹¹⁸, P[w⁺, fat facets-lacZ]/Y males were mutagenized using 25 mM ethyl methane sulfonate (EMS) in 1% sucrose (LEWIS and BACHER 1968 ) and crossed to y w hnt¹¹⁴², P[ry⁺, FRT¹⁰¹]/FM7, Df(1)KA14/FM7c, or Df(1)HA32/FM7c, P[ry^+t7.2, ftz-lacZ] (FM7Z) virgin females. After 4 days, the adults were cleared from the bottles. The resulting female offspring carrying the balancer chromosome were then individually crossed to FM7/Y, FM7c/Y, or FM7Z/Y males to establish stocks. The stocks were then expanded, and the embryos were collected using 50-ml conical tubes with apple juice-agar medium in the caps. Chorion removal, fixation, and staining were carried out in 48-well custom staining trays. The germ cells were labeled using X-Gal as the substrate (SIMON et al. 1985 ; HOLMES et al. 1998 ).

View larger version (17K): In this window In a new window Download PPT slide   Figure 2. Diagram of a screen for X-linked mutations affecting germ cell development. Males were mutagenized using EMS. The mutagenized X chromosome (*) was isolated over an FM7 balancer chromosome. Depending upon the viability and fertility of the mutagenized X chromosome, either homozygous or balanced stocks were established. Overnight collections of embryos from these stocks were used for staining.

The nonmutagenized w¹¹¹⁸, P[w⁺, fat facets-lacZ] parental stock had a very low percentage of embryos with any germ cells outside the gonad. During the screen and subsequent analyses, embryos were scored as mutant if they had four or more germ cells ectopic to the gonad. For brevity, the parental w¹¹¹⁸, P[w⁺, fat facets-lacZ], mutagenized w¹¹¹⁸ sctt, P[w⁺, fat facets-lacZ], and w¹¹¹⁸ out, P[w⁺, fat facets-lacZ], chromosomes will be referred to as faf-lacZ, sctt, and out, respectively.

Mapping and complementation analyses:
To determine the location of the sctt and out mutations, the mutant lines were tested for complementation using the X chromosome deletion stocks available from the Bloomington and Umeå stock centers (FLYBASE 1999 ). Since sctt is required maternally, sctt/Df(1) females were generated and their offspring were tested for abnormal germ cell development. The location of out was established by crossing Df(1)/balancer virgin females to out/Y males. The germ cell phenotype of the progeny of the Df(1)/out females was then assayed.

Recombination mapping to localize sctt was performed using the sctt mutant chromosome and w¹¹¹⁸ cv¹ wy⁷⁴ⁱ f¹, sc¹ ec¹ cv¹ wy⁷⁴ⁱ f¹, or y¹ w¹ cv¹ ct⁶ v¹ marker chromosomes. The insertion site of the P[w⁺, fat facets-lacZ] was determined to be 18F-19A by in situ hybridization of a digoxygenin-labeled white probe (a kind gift from Joseph Heilig) to salivary gland chromosomes (O'HARE et al. 1984 ; JOHNSON-SCHLITZ and LIM 1987 ; TAUTZ and PFEIFLE 1989 ; KLINGLER and GERGEN 1993 ). Virgin females homozygous for the sctt mutant chromosome were crossed to males carrying the marker X chromosome. Recombination occurred in the F₁ females, which were crossed to FM7c/Y males. Individual F₂ males from this cross that carried a recombinant chromosome were then crossed to y w hnt¹¹⁴², P[ry⁺, FRT¹⁰¹]/FM7c, Df(1)KA14/FM7c, or Df(1)HA32/FM7Z virgin females. Each individual recombinant X chromosome was then isolated over a balancer chromosome to establish a stock. The balancer chromosome was removed to yield an independent homozygous recombinant line. Embryos collected from F₅ or later mothers homozygous for the recombinant X chromosome were scored for the sctt phenotype using X-Gal staining of 12- to 15-hr embryos.

Complementation tests between sctt, out, and the other lines obtained in the screen were also performed. To test for complementation of the maternal-effect germ cell migration phenotype of sctt, virgin females from the remaining lines were crossed to sctt/Y males. The offspring from trans-heterozygous sctt/test chromosome F₁ females were then assayed for complementation. The out complementation group was determined by testing out/FM7Z and/or out/out females crossed to males from the other lines, if the test line produced viable males. If the X chromosome carried a lethal mutation, heterozygous test chromosome/FM7Z virgin females were crossed to out/Y males. The germ cell phenotype of the progeny was then assayed by ß-galactosidase (ß-Gal) staining.

Fertility assay and ovary dissections:
To test fertility, individual male and female offspring of sctt/sctt, sctt/FM7, or faf-lacZ/faf-lacZ females were crossed to Ore-R flies and placed in vials for 5–7 days. For the crosses testing male fertility, vials were scored for the presence of larvae or pupae. A vial that contained no offspring was scored as sterile only if adults of both sexes were still alive in the vial and unhatched eggs were present. In the crosses that tested female fertility, vials were scored for the presence of eggs or offspring. In these cases, a vial that did not contain eggs was scored as sterile only if the test female was still alive at the end of the trial period.

Ovaries from some of the test females were dissected in PBST (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, 0.3% Triton X-100, pH 7.2; SAMBROOK et al. 1989 ) to assay the status of ovary development. Ovaries containing late-stage oocytes were scored as normal (KING 1970 ; MAHOWALD and KAMBYSELLIS 1980 ; SPRADLING 1993 ). Agametic ovaries had no detectable oocytes. Rudimentary ovaries contained early oogenic stages (younger than stage 9), but no late-stage oocytes. The data presented include only those females from which both ovaries were recovered in the dissection.

Germ cell counts:
To determine the number of ß-Gal-positive germ cells present in 11- to 12-hr faf-lacZ control, sctt, or out embryos, embryos were collected for 1 hr on apple juice caps and then aged 11 hr at 25°. ß-Gal staining was performed as above and germ cells were counted using differential interference contrast microscopy. All of the cells counted were positive for ß-Gal activity and exhibited the spherical nucleus and large cell size characteristic of germ cells at this stage of development (RABINOWITZ 1941 ; POULSON 1950 ; UNDERWOOD et al. 1980 ). The gonadal sheath cells were used to determine the boundary of the gonad. Statistical analyses were performed using Minitab Statistical Software (Minitab, State College, PA) or JMP version 4.0 (SAS Institute, Cary, NC).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Screen for mutants defective in germ cell development:
X-linked genes with functions required for germ cell development were identified in an EMS mutagenesis screen of 3475 independent lines. Germ cells ectopic to the gonad were observed in 39 lines. Seventeen of these lines exhibited >40% penetrance and were kept for further analyses. In the four most severe mutants, the germ cells are scattered throughout the posterior half of the embryo and very few, if any, germ cells populate the gonad. This scattered germ cell phenotype is observed in two homozygous viable lines and two lines that are homozygous lethal. A second class of 13 lines exhibit less severe phenotypes that are characterized by visible clusters of germ cells in the position of the gonad and four or more germ cells located outside of the gonads. This group consists of eight homozygous viable alleles and five chromosomes with lethal mutations. Complementation analyses of these 17 lines reveal that the mutations we isolated represent at least four complementation groups. Detailed analyses of the sctt and out genes are described here.

Genetic characterization of sctt:

A single recessive allele of sctt was isolated (Fig 3). This allele is highly penetrant (>94%), and disruption of the germ cell staining pattern requires that the mother be homozygous mutant (Table 1). The offspring of sctt/sctt females will be referred to as sctt mutant embryos unless otherwise specified. This allele is complemented by all of the X chromosome deletion stocks available from the Bloomington and Umeå stock centers, which represent 80% of the X chromosome. Since a noncomplementing deletion was unavailable, we determined the genetic map position by recombination mapping of sctt using marked X chromosomes. We analyzed 32 lines with a recombination event between cv and wy and 21 lines with a recombination event between y or sc and cv. These analyses place sctt within 1 map unit of cv. The cv gene maps to 1-13.7 on the genetic map and is predicted to lie within polytene chromosome bands 5A13–5B1 (LINDSLEY and ZIMM 1992 ; FLYBASE 1999 ). The Dp(1:Y)dx⁺5 chromosome carrying polytene chromosome bands 4C11;6D8 zygotically rescues the sctt germ cell migration phenotype. This indicates that sctt is within the 4C11:6D8 interval. Three gaps in the deletion stocks fall within the region delimited by our recombination mapping. It is likely that sctt is located in one of these regions not represented by the deletion stocks. It is formally possible that the sctt phenotype is the result of two independent mutations. If this is the case, they are very tightly linked. The sctt phenotype is complemented by all P-element insertions and previously identified mutations tested to date.

View larger version (38K): In this window In a new window Download PPT slide   Figure 3. Mutation of sctt results in aberrant germ cell migration and retention of germ cell marker expression in germ cells outside of the gonad. Lateral views of (A) sctt mutant and (C) wild-type embryos showing germ cells migrating across the posterior midgut epithelium (arrows). Germ cell formation and migration are indistinguishable from a wild-type embryo. Dorsal views of stage 14 sctt (B) and wild-type (D) embryos are shown. Note that very few germ cells, if any, in the sctt embryo are located in the regions of the gonad (arrows). Also, many germ cells outside of the gonad continue to express Faf-ß-Gal. The embryos shown in A–D were stained in parallel. (E) An sctt mutant embryo double labeled for the 412 retrotransposon, a marker of somatic gonadal mesoderm, and for the Faf-ß-Gal germ cell marker. The somatic gonadal mesoderm (arrows, brown) shows a wild-type staining pattern. The germ cells are blue and are seen scattered throughout the posterior of the embryo.

Genetic characterization of out:

The out complementation group consists of six recessive alleles (Table 2). In out mutants, clusters of Faf-ß-Gal-expressing germ cells are clearly localized to each gonad, but germ cells ectopic to the gonad are also prominent (Fig 4). Three alleles are homozygous viable (out¹, out², and out³), and three alleles are homozygous lethal (out⁴, out⁵, and out⁶). The product of the out gene is required zygotically. Penetrance of the mutant phenotype ranges from 65.6% for out⁶ to 93.7% for out¹. The three viable alleles, out¹, out², and out³, fail to complement the germ cell mutant phenotype of each other and, for each of these alleles, this phenotype is uncovered by Df(1)JA27, which is reported to be a deletion of polytene chromosome bands 18A5;18D1 (FLYBASE 1999 ). The three viable alleles fail to complement out⁴ and out⁵. Interestingly, out⁶ is not complemented by out¹ and out³ but is complemented by out² (Table 2). Embryos from the three lethal lines appear to have normal segmentation; however, the out⁴ and out⁵ lines exhibit increased pupal lethality. The phenotypes of Df(1)JA27/Y males and Df(1)JA27/out¹ hemizygous females are more severe than any combination of out alleles from our screen. The germ cell phenotype associated with the Df(1)JA27 chromosome is recessive, as Df(1)JA27/+ embryos have a wild-type germ cell staining pattern. The out phenotype is complemented by all P-element and known mutations within the 18A5;18D1 interval that we have tested. We have not determined whether the lethality associated with out⁴, out⁵, and out⁶ is due to the out mutation or is due to a second mutation present on these chromosomes.

View larger version (112K): In this window In a new window Download PPT slide   Figure 4. The out mutants are deficient in the elimination of ectopic germ cells. (A) A dorsal view of a stage 14 out/out mutant embryo shows two gonads. However, many cells outside of the gonad continue to stain for Faf-ß-Gal. (B) Dorsal view of a wild-type embryo stained in parallel with the embryo shown in A.

Mutation of sctt disrupts germ cell migration and developmentally regulated cell death:
In sctt mutant embryos, germ cell development, as visualized by the Faf-ß-Gal marker, appears normal prior to migration. Germ cells form at the posterior pole, and they are incorporated into the posterior midgut (PMG) pocket at the completion of germ band extension (stage 10; CAMPOS-ORTEGA and HARTENSTEIN 1997 ). The germ cells then transit the epithelium of the PMG (JAGLARZ and HOWARD 1994 , JAGLARZ and HOWARD 1995 ; WARRIOR 1994 ; CALLAINI et al. 1995 ; Fig 3A). However, very few, if any, of the germ cells in sctt mutant embryos successfully migrate to and become incorporated into the developing gonad. Instead, the germ cells disperse throughout the posterior half of the embryo and continue to express the Faf-ß-Gal marker (compare Fig 3B and Fig D, and Fig 5A). This phenotype reveals that this sctt mutation disrupts two processes. First, cell migration is severely impaired. Second, the cells that do not reach the gonad fail to undergo cell death. Despite the severity of the germ cell phenotype, homozygous sctt mutants are viable, and these individuals do not display any detectable morphological abnormalities as embryos or adults, with the exception of agametic gonads (see below).

View larger version (80K): In this window In a new window Download PPT slide   Figure 5. The sctt germ cell migration phenotype can be paternally rescued. Zygotic expression of a wild-type copy of sctt rescues the germ cell migration defect, but cells outside of the gonad continue to express germ cell markers. Two embryos from sctt/sctt mutant mothers are shown. (A) The phenotype observed in homozygous sctt/sctt or hemizygous sctt/Y individuals. (B) An sctt/+ embryo. Note the small clusters of germ cells that have successfully migrated to the gonad (arrows in B).

In the homozygous sctt stock, 97.8% of the embryos display the extremely severe phenotype shown in Fig 3B. In contrast, only 3–4.5% of embryos from sctt/FM7 mothers and 6.5% of embryos from the unmutagenized faf-lacZ/faf-lacZ stock have four or more labeled germ cells located outside of the gonad (Table 1). The germ cell migration defect is observed only among the offspring of homozygous sctt mutant mothers, and it is not evident in the sctt/sctt or sctt/Y offspring of sctt/FM7 females crossed to sctt/Y males. Therefore, the sctt mutation isolated in this screen is recessive and a maternal effect. Also, the phenotypes of sctt/sctt and sctt/Y embryos from sctt/sctt mutant mothers are equivalent.

The pathfinding and cell death defects in sctt embryos can be genetically uncoupled:

To better understand the relationship between germ cell migration and programmed cell death, we performed additional genetic and phenotypic analyses of sctt mutant embryos. We reasoned that if sctt function is required at different times for either germ cell migration or the elimination of ectopic germ cells, then we might be able to separate differing temporal requirements through paternal contribution of a wild-type copy of sctt. If sctt function were required both maternally and zygotically for germ cell migration, then an embryo expressing a wild-type copy of sctt would display a less severe germ cell migration phenotype. Alternatively, if the cell death phenotype were rescued by zygotic expression of a wild-type copy of sctt, then only those germ cells that successfully migrated to the gonad would continue to express Faf-ß-Gal, while those cells outside of the gonad would undergo programmed cell death.

When sctt/sctt females are crossed to Ore-R males, the offspring of this cross fall into two phenotypic classes (Table 1, line 2; Fig 5). The first class displays the severe germ cell migration defect seen in the homozygous sctt line, and the germ cells outside the gonad continue to express Faf-ß-Gal (Fig 5A). In these embryos, so few germ cells successfully migrate that the gonads cannot be easily identified in a ß-Gal-stained embryo. This phenotype was observed in 60% (n = 653) of the embryos. In the remaining embryos, an increased number of germ cells successfully migrate to the gonad, but many cells outside of the gonad continue to express Faf-ß-Gal (Fig 5B). To test the hypothesis that the less severe phenotype is the result of paternal rescue, we crossed sctt/sctt females to FM7Z/Y males. This hypothesis predicts that the more severe phenotype will be observed in sctt/Y embryos, while the less severe phenotype will be observed in sctt/FM7Z embryos. The FM7Z chromosome carries a wild-type sctt gene. The sctt/FM7Z female embryos resulting from this cross stain blue due to the presence of the ftz-lacZ transgene. The severe sctt phenotype was observed in 217/218 sctt/Y embryos (Table 1, line 3). Among the lightly stained sctt/FM7Z female embryos, ß-Gal-stained germ cells within the gonad as well as germ cells located outside of the gonad were observed. Therefore, zygotic expression of wild-type sctt from the paternal X chromosome improves germ cell migration. The recessive maternal effect, the recessive zygotic phenotypes, and the zygotic rescue data collectively suggest that this sctt allele is loss of function.

We counted the number of germ cells that successfully migrated to the gonad in embryos that were born to sctt/sctt mothers but received a wild-type paternal copy of sctt. Interestingly, the same numbers of germ cells were observed within the gonads of control and paternally rescued embryos at 11–12 hr of development (Table 3). We counted an average of 13.7 germ cells within the two gonads of faf-lacZ embryos (n = 50) and 12.7 germ cells incorporated into the two gonads of sctt/+ rescued embryos (n = 50). These values are not statistically different (P = 0.09, Student's t-test). The average of 6–7 germ cells per gonad is within the range reported for other Drosophila lines (SONNENBLICK 1941 , SONNENBLICK 1950 ; UNDERWOOD et al. 1980 ; HAY et al. 1988 ). These data support the conclusion that zygotic expression of a wild-type copy of sctt rescues the germ cell migration defect associated with sctt mutant embryos. However, an average of 7.7 germ cells was observed outside of the gonads in the rescued embryos. The range was 4–16 germ cells ectopic to the gonad, and most of the Faf-ß-Gal positive cells retained a characteristic germ cell morphology (RABINOWITZ 1941 ; POULSON 1950 ; UNDERWOOD et al. 1980 ). Thus, the developmental cell death of ectopic germ cells is not rescued by zygotic expression of wild-type sctt. The average total number of germ cells observed is 17.7 (n = 50) in sctt/sctt or sctt/Y embryos and 20.4 in sctt/+ embryos (n = 50; Table 3). Both of these means are significantly higher than the total germ cell number of 14.0 observed in controls (P 0.0001, Student's t-test).

Fertility is greatly reduced in sctt mutants:
The sctt phenotype cosegregates with reduced fecundity. To quantitate this phenotype, we performed fertility tests. Male and female offspring of faf-lacZ/faf-lacZ control, sctt/FM7, and sctt/sctt females were assayed for fertility. When individual test males or females from faf-lacZ/ faf-lacZ or sctt/FM7 mothers were crossed to Ore-R flies of the opposite sex, the fertility rate was >90% (Table 4). However, sctt males from sctt/sctt mutant mothers were fertile in only 33.7% of the crosses (n = 285). When sctt/sctt females from sctt/sctt mutant mothers were mated to Ore-R males, 41.7% of the females produced offspring (n = 163). The remaining females failed to lay any eggs. Interestingly, sctt/+ females from sctt/sctt mothers were as fertile as wild-type animals (100%, n = 135). Therefore, the zygotic rescue of the germ-cell-migration defect associated with the sctt mutation restores fertility to wild-type levels. This is consistent with the observation that wild-type numbers of germ cells are incorporated into the gonad of these paternally rescued embryos. The ovaries from these individuals were also normal (see below).

Agametic ovaries occur at a high frequency in sctt mutants:
The sterility associated with the sctt mutation suggested a lack of germ-line stem cells within the gonad, an arrest of oogenesis, or abnormal somatic gonad development. We investigated these possibilities by analyzing ovarian development in different mutant backgrounds. An absence of germ-line stem cells or an early disruption of oogenesis would result in agametic ovaries, while abnormal somatic gonad development would lead to a lack of or malformation of somatically derived ovarian structures.

The formation of a functional gonad requires the interaction of the germ cells with the somatic gonadal mesoderm. To test whether the somatic gonad cell fate is specified in sctt mutant embryos, we used the 412 retrotransposon as a molecular marker for somatic gonad differentiation. The 412 retrotransposon probe labels gonadal mesoderm cells even in the absence of germ cells (BROOKMAN 1992). The in situ hybridization staining pattern in sctt mutant embryos is indistinguishable from wild type (Fig 3E). In addition, the fact that sctt embryos produce viable and fertile adults, albeit at a lower frequency, demonstrates that the gonad can function if populated by germ cells.

The dissection of ovaries from control and sctt mutants reveals that the sterility associated with the sctt mutation is the result of a lack of germ-line stem cells or a very early arrest of oogenesis. Of the ovaries dissected from 87 offspring of faf-lacZ/faf-lacZ or sctt/FM7 females, 99% were normal (n = 174 ovaries; Table 5; Fig 6B). Two ovaries from the sctt/FM7 test lines were scored as rudimentary because they did not contain oocytes older than stage 9 (KING 1970 ; SPRADLING 1993 ). In contrast, among the sctt/sctt female offspring of sctt/sctt mutant mothers, 57% of the ovaries were agametic, 11% were rudimentary, and 32% were normal (n = 122). An example of agametic ovaries from an sctt/sctt mutant is shown in Fig 6A. The gonadal defects correlate with fecundity. We dissected 32 females that did not lay eggs. Of these, 24 had two agametic ovaries, 6 had two rudimentary ovaries, and 2 had one agametic ovary and one rudimentary ovary. We examined an additional 29 sctt females that did lay eggs. All of these had at least one normal ovary and, in 10 of them, both ovaries appeared normal. Interestingly, 97% (n = 120) of the ovaries from sctt/+ females from sctt/sctt mothers were wild type. Therefore, zygotic rescue of the germ cell migration defect leads to nearly normal ovary formation and restoration of wild-type fertility levels.

View larger version (73K): In this window In a new window Download PPT slide   Figure 6. The sterility associated with the sctt mutation is due to an absence of germ cells within the gonad. (A) Agametic ovaries from an sctt mutant have a normal complement of somatic gonad-derived structures, but lack germ cells. (B) Wild-type ovaries. Both photographs are taken at the same magnification.

The out mutation impairs cell death of errant germ cells:
An example of an out¹/out¹ mutant embryo is shown in Fig 4A. Somatic gonad development as assayed by the 412 retrotransposon is normal (data not shown). The out gene product is required zygotically, and the six out alleles result in many ß-Gal-labeled germ cells outside of the gonad. These ß-Gal-positive cells can be seen outside of the gonad until at least 15 hr of development. At this time, cuticle secretion inhibits effective staining (CAMPOS-ORTEGA and HARTENSTEIN 1997 ). Thus, errant germ cells persist in out mutants and continue to express the faf-lacZ marker.

To determine whether this defect in cell death is linked to a migration defect, we determined the number and location of germ cells present in 11- to 12-hr embryos. The results are presented in Table 3. Control embryos had an average of 13.7 germ cells (n = 50) in the gonads compared to an average of 13.0 germ cells in out¹/out¹ mutants (n = 50). These numbers are not statistically different (P = 0.16, Student's t-test). An average of 8.6 germ cells was observed ectopic to the gonad in the mutants, compared to 0.3 ß-Gal-positive germ cells outside of the gonads in controls. The range was 4–18 ectopic germ cells in the out mutants. The average total number of germ cells in out¹/out¹ embryos is 21.6. This number is statistically different from the average of 14.0 observed in controls (P < 0.0001, Student's t-test).

As an additional test of whether the Faf-ß-Gal-positive cells outside of the gonad are continuing to express the germ cell fate, we analyzed the expression of Vasa-GFP, another germ cell-specific marker (BREITWIESER et al. 1996 ). The Vasa-GFP germ cell labeling pattern is indistinguishable from the Faf-ß-Gal staining pattern in out¹/out¹ embryos (data not shown). In addition, the Faf-ß-Gal-labeled cells are larger than neighboring cells and possess the distinct spherical nucleus typical of germ cells (RABINOWITZ 1941 ; POULSON 1950 ; UNDERWOOD et al. 1980 ). Therefore, the mislocalized cells in out/out embryos continue to display three independent germ cell traits. These data suggest that mutations of the out gene disrupt the programmed cell death of the germ cells.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We report the identification of two X-linked genes, sctt and out. When mutated, sctt severely disrupts germ cell migration, and the germ cells ectopic to the gonad fail to undergo programmed cell death. The sctt mutant flies are often sterile, and the females have a high percentage of agametic ovaries. Zygotic expression of a wild-type copy of sctt rescues germ cell migration, but not the germ cell death defect. Thus, the sctt mutation disrupts at least two developmental processes: the pathfinding mechanisms that enable the germ cells to migrate to the gonad and the regulation of cell survival. In out mutant embryos, the same numbers of germ cells are incorporated into the gonads as in wild-type embryos. Thus, germ cell migration is normal. However, germ cells ectopic to the gonads do not undergo programmed cell death. We propose two models for out function. The function of out may be part of a cell death mechanism responsible for the elimination of germ cells ectopic to the gonad. Alternatively, a mutation in out may result in increased activity of a survival factor that allows germ cells ectopic to the gonad to persist.

Maternal sctt regulates cell migration:
The sctt phenotype has several noteworthy features. First, it requires that the mother be homozygous mutant. To date, only the products of the nos and pgc-1 genes are known to be required in the germ cells for normal migration (KOBAYASHI et al. 1996 ; NAKAMURA et al. 1996 ; FORBES and LEHMANN 1998 ). Both are maternally contributed. Mutation of pgc-1, an untranslatable RNA, results in decreased nos RNA levels after the cellular blastoderm stage. Therefore, the germ cell phenotype of pgc-1 embryos may be the result of its effect on nos expression. Nos, a zinc-finger protein, is necessary to suppress transcription and mitosis in the germ cells (KOBAYASHI et al. 1996 ; ASAOKA et al. 1998 ; FORBES and LEHMANN 1998 ; DESHPANDE et al. 1999 ). Zygotic expression of nos does not rescue the germ cell migration defect (FORBES and LEHMANN 1998 ). Since the germ cell migration defect of sctt mutants can be rescued zygotically, sctt may function in germ cell migration later than nos. In fact, the nos germ cell migration phenotype is clearly distinct from the sctt phenotype. Germ cells deprived of maternal nos expression exit the posterior midgut and then cluster close to the posterior midgut rudiment (FORBES and LEHMANN 1998 ). The sctt mutant germ cells cross the midgut epithelium and disperse individually throughout the posterior of the embryo.

Since the germ cell migration phenotype of sctt mutant embryos is so severe, we favor the hypothesis that the agametic ovaries are caused by a failure of germ-line stem cells to populate the gonads. A similar agametic ovary phenotype has been reported for wunen mutants (ZHANG et al. 1996 ). Paternal rescue of germ cell migration and the concomitant restoration of fertility by a wild-type copy of sctt are consistent with this interpretation. However, it is possible that the sctt mutation disrupts a later developmental event within the ovary, resulting in a degeneration of germ-line stem cells or a very early arrest of oogenesis.

It is likely that additional alleles of sctt were generated in our screen, but they were not successfully retained. To observe the germ cell defects resulting from a mutation of sctt, a homozygous mutant female must be viable and fertile. Therefore, lethal or sterile chromosomes would not have been recovered. In addition, sctt males are only 33.7% fertile, and females are only 41.7% fertile. Homozygous viable lines may have been lost because of a failure to reproduce. Noncomplementation screens specifically designed to isolate additional alleles of sctt will be required for a more thorough genetic analysis of the sctt locus.

Cell migration coupled to cell survival:
Another interesting feature of sctt is that it affects both germ cell migration and germ cell survival. However, the temporal requirements for sctt in these two processes are different. The effect on germ cell death is not rescued by zygotic expression of a wild-type copy of the gene. In some experimental systems, migration guidance cues have been shown to be integrally linked to or to have additional functions as survival factors (FLEISCHMAN 1993 ; GOETZL and AN 1998 ; KUNISADA et al. 1998 ; WAKAMATSU et al. 1998 ; ASHMAN 1999 ; DE FELICI 2000 ; GOETZL et al. 2000 ; SWARTHOUT and WALLING 2000 ; FUKUSHIMA et al. 2001 ; KIERSZENBAUM and TRES 2001 ; MILLER and KAPLAN 2001 ). For example, in avian and mouse embryos, the receptor tyrosine kinase, c-Kit, and its ligand, Stem cell factor, have been shown to act as both cell migration and anti-apoptotic factors in both neural crest and germ cell lineages (FLEISCHMAN 1993 ; KUNISADA et al. 1998 ; WAKAMATSU et al. 1998 ; ASHMAN 1999 ; DE FELICI 2000 ; KIERSZENBAUM and TRES 2001 ). In neurons, neurotrophins can act as both prosurvival and proapoptotic signals, depending upon the developmental context and receptor configuration of the receiving cell (MILLER and KAPLAN 2001 ).

Recently, DESHPANDE et al. 2001 demonstrated that ectopic expression of hedgehog could act as an attractive signal for Drosophila germ cells. In avian embryos, CHARRIER et al. 2001 have shown that Sonic hedgehog can act as an anti-apoptotic factor in neural tube formation. Considering the proposed role of Hedgehog as an attractive signal for germ cell migration in Drosophila, it is possible that Hedgehog may be involved in both pathfinding and cell survival in germ cell development.

The lysophospholipid class of molecules, including lysophosphatidic acid and sphingosine 1-phosphate, has been shown to mediate both survival and cell migration behaviors (GOETZL and AN 1998 ; GOETZL et al. 2000 ; SWARTHOUT and WALLING 2000 ; FUKUSHIMA et al. 2001 ). In zebrafish, a sphingosine-1-phosphate receptor has been shown to be necessary for heart precursor cell migration (KUPPERMAN et al. 2000 ). The recently identified roles of various lipid-modifying enzymes in Drosophila germ cell migration, including a phosphatidic acid phosphatase and 3-hydroxy-3-methylglutaryl coenzymeA reductase, make lipid-based signals likely candidates for germ cell guidance cues (ZHANG et al. 1996 , ZHANG et al. 1997 ; VAN DOREN et al. 1998 ; STARZ-GAIANO et al. 2001 ). Overexpression of wunen can act at a distance to decrease germ cell number (STARZ-GAIANO et al. 2001 ). Deciphering the potential role of the lysophospholipid class of signaling molecules in cell migration and/or cell survival will be a particularly interesting area of study.

It is possible that inappropriate cell division may contribute to the sctt and out cell death phenotypes. In nos mutant germ cells and patched mutant embryos, germ cells reenter the cell cycle prematurely and the number of germ cells is increased (DESHPANDE et al. 1999 , DESHPANDE et al. 2001 ). Germ cells ectopic to the gonad are frequently observed in these embryos. However, the total number of germ cells we observe in out and sctt mutant embryos is less than that observed in either the nos or the patched mutant backgrounds. Other observations suggest that the extra germ cells observed in the sctt and out mutants are insufficient to override the embryo's ability to eliminate these additional cells. When a 39% increase in germ cell number is caused by the overexpression of osk in 6X osk embryos, ectopic germ cells are not seen after gastrulation (SMITH et al. 1992 ). In our faf-lacZ genetic background, a mutation that causes a 29% increase in the number of germ cells forming at the posterior pole does not result in an increase in germ cells ectopic to the gonads (C. R. COFFMAN and R. E. BOSWELL, unpublished results). Therefore, the Drosophila embryo appears to have some additional capacity for destroying germ cells ectopic to the gonads, and the generation of additional germ cells cannot completely account for the out and sctt phenotypes.

Programmed cell death and Drosophila germ cells:
There are multiple types of programmed cell death. Programmed cell death via apoptosis and autophagy has been reported in Drosophila (ABRAMS et al. 1993 ; FOLEY and COOLEY 1998 ; ABRAMS 1999 ; BANGS and WHITE 2000 ; LEE and BAEHRECKE 2000 , LEE and BAEHRECKE 2001 ). While apoptosis and autophagy have distinct morphologies, they utilize some of the same molecular machinery and therefore cannot always be distinguished using molecular markers.

Over 60 years ago, Drosophila biologists demonstrated that the number of primordial germ cells exceeds the number of germ cells incorporated into the gonads (RABINOWITZ 1941 ; SONNENBLICK 1941 ). These authors and others have suggested that this reduction in germ cell number is likely to occur via an active and highly regulated process (SONNENBLICK 1950 ; TECHNAU and CAMPOS-ORTEGA 1986 ; STARZ-GAIANO et al. 2001 ). However, a direct demonstration of the cellular mechanism or mechanisms responsible for the elimination of these germ cells has not been reported. STARZ-GAIANO et al. (2001) make the interesting observation that forced expression of Wrinkled [W, a.k.a. head involution defective (hid)] or reaper (rpr) in germ cells results in their death. This argues that the machinery for apoptotic or autophagic cell death is present in the germ cells, as both of these programmed cell death mechanisms utilize W/hid or rpr functions (LEE and BAEHRECKE 2001 ; THUMMEL 2001 ).

High levels of p53 RNA expression are observed in Drosophila primordial germ cells. This may reflect a function for this pro-apoptotic gene in the elimination of germ cells that have experienced DNA damage. In support of this hypothesis, p53 has been shown to activate the expression of rpr, a cell death activator, in cells exposed to radiation-induced DNA damage (BRODSKY et al. 2000 ; OLLMANN et al. 2000 ). Determining whether the elimination of germ cells ectopic to the gonad utilizes the caspase-mediated pathways activated by the cell death inducers rpr, grim, sickle, and W (ABRAMS et al. 1993 ; ABRAMS 1999 ; BANGS and WHITE 2000 ; LEE and BAEHRECKE 2000 , LEE and BAEHRECKE 2001 ; CHRISTICH et al. 2002 ; SRINIVASULA et al. 2002 ) or by some alternative cell death pathway such as the one reported for nurse cells (FOLEY and COOLEY 1998 ), will require further investigation. Future studies may also reveal roles for genes that affect the expression and/or regulation of cell death inhibitors such as iap1/thread, iap2, or Deterin (ABRAMS 1999 ; VERNOOY et al. 2000 ). None of the known cell death genes identified to date map to the sctt or out regions.

The presence of a large number of germ cells outside of the gonads in both sctt and out mutants indicates that mutations in these genes affect the elimination of these germ cells and that the death of germ cells ectopic to the gonads occurs via an active, developmentally regulated process. Both phenotypes are recessive, and in the case of out, germ cells ectopic to the gonad are observed when out is placed over a deletion. This indicates that these out alleles are loss of function. However, we have not demonstrated whether apoptosis, autophagy, or some other mechanism of programmed death is disrupted in sctt and out mutants. We also cannot rule out the formal possibility that necrosis is responsible for some germ cell loss during development. It has been proposed that the lost germ cells may lack sufficient germ plasm (SONNENBLICK 1950 ).

The germ cells of Drosophila are an excellent model for studying cell death, and there is great potential for identifying additional genes necessary for this process. Since only 50% of the germ cells successfully reach the gonad, 10–30 germ cells normally undergo developmental cell death in the embryo (SONNENBLICK 1941 , SONNENBLICK 1950 ; UNDERWOOD et al. 1980 ; TECHNAU and CAMPOS-ORTEGA 1986 ).

Mutations that affected germ cell survival occurred very frequently in our screen. There are at least two explanations for this. First, the ability to detect germ cells outside the gonad is requisite for a screen like the one reported here. If the "lost" germ cells immediately ceased expression of germ cell-specific markers, then the phenotype of a germ cell migration defective mutant would be a reduction in the number of germ cells within the gonad and might go undetected. Therefore, our screen was very effective at identifying mutations that affect germ cell death as well as germ cell migration.

Second, the mechanisms regulating germ cell survival and death are likely to have many components. Therefore, there are many potential targets for mutagenesis screens. Since very few genes have been identified with roles in Drosophila germ cell migration and death, additional screens will be required to achieve a more complete understanding of these processes. In addition, it will be interesting to examine potential cell death defects in other germ cell migration mutants identified and to investigate possible interactions between these genes and sctt and/or out. Further, it will be important to determine whether sctt and out function in the same pathway or in distinct pathways to regulate germ cell survival.

	ACKNOWLEDGMENTS

We thank Janice Fischer, Joseph Heilig, Linda Ambrosio, Joseph Duffy, and the Bloomington and Umeå stock centers for providing the Drosophila stocks used in this work. Jo Anne Powell-Coffman, Kristen Johansen, William Nordstrom, Angela Mortvedt, and Corinne Glynn provided critical comments on the manuscript. C.R.C. was supported by a Howard Hughes Medical Institute postdoctoral fellowship at the University of Colorado where the initial phases of this work were performed. The genetic and phenotypic analyses of out and sctt were conducted at Iowa State University and were supported by funds provided to C.R.C. by the Iowa Agriculture and Home Economics Experiment Station. R.E.B. is funded by National Institutes of Health grant GM-57989. This manuscript, designated as J-19748 of the Iowa Agriculture and Home Economics Experiment Station, Ames, Iowa, project no. HIOW03383, was supported by Hatch Act and State of Iowa funds.

Manuscript received February 21, 2002; Accepted for publication June 21, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABRAMS, J. M., 1999  An emerging blueprint for apoptosis in Drosophila. Trends Cell Biol. 9:435-440.[Medline]

ABRAMS, J. M., K. WHITE, L. I. FESSLER, and H. STELLER, 1993  Programmed cell death during Drosophila embryogenesis. Development 117:29-43.[Abstract]

ASAOKA, M., H. SANO, Y. OBARA, and S. KOBAYASHI, 1998  Maternal Nanos regulates zygotic gene expression in germline progenitors of Drosophila melanogaster.. Mech. Dev. 78:153-158.[Medline]

ASHMAN, L. K., 1999  The biology of stem cell factor and its receptor C-kit. Int. J. Biochem. Cell Biol. 31:1037-1051.[Medline]

BANGS, P. and K. WHITE, 2000  Regulation and execution of apoptosis during Drosophila development. Dev. Dyn. 218:68-79.[Medline]

BOYLE, M., N. BONINI, and S. DINARDO, 1997  Expression and function of clift in the development of somatic gonadal precursors within the Drosophila mesoderm. Development 124:971-982.[Abstract]

BREITWIESER, W., F. H. MARKUSSEN, H. HORSTMANN, and A. EPHRUSSI, 1996  Oskar protein interaction with Vasa represents an essential step in polar granule assembly. Genes Dev. 10:2179-2188.[Abstract/Free Full Text]

BRODSKY, M. H., W. NORDSTROM, G. TSANG, E. KWAN, and G. M. RUBIN et al., 2000  Drosophila p53 binds a damage response element at the reaper locus. Cell 101:103-113.[Medline]

BROIHIER, H. T., L. A. MOORE, M. VAN DOREN, S. NEWMAN, and R. LEHMANN, 1998  zfh-1 is required for germ cell migration and gonadal mesoderm development in Drosophila. Development 125:655-666.[Abstract]

BROOKMAN, J. J., A. T. TOOSY, L. S. SHASHIDHARA, and R. A. WHITE, 1992  The 412 retrotransposon and the development of gonadal mesoderm in Drosophila. Development 116:1185-1192.[Abstract]

CALLAINI, G., M. G. RIPARBELLI, and R. DALLAI, 1995  Pole cell migration through the gut wall of the Drosophila embryo: analysis of cell interactions. Dev. Biol. 170:365-375.[Medline]

CAMPOS-ORTEGA, J., and V. HARTENSTEIN, 1997 The Embryonic Development of Drosophila melanogaster. Springer-Verlag, New York.

CHARRIER, J. B., F. LAPOINTE, N. M. DOUARIN, and M. A. TEILLET, 2001  Anti-apoptotic role of Sonic hedgehog protein at the early stages of nervous system organogenesis. Development 128:4011-4020.[Abstract/Free Full Text]

CHRISTICH, A., S. KAUPPILA, P. CHEN, N. SOGAME, and S. I. HO et al., 2002  The damage-responsive Drosophila gene sickle encodes a novel IAP binding protein similar to but distinct from reaper, grim, and hid.. Curr. Biol. 12:137-140.[Medline]

COUNCE, S. J., 1963  Developmental morphology of polar granules in Drosophila including observations on pole cell behavior and distribution during embryogenesis. J. Morphol. 112:129-145.

DE FELICI, M., 2000  Regulation of primordial germ cell development in the mouse. Int. J. Dev. Biol. 44:575-580.[Medline]

DESHPANDE, G., G. CALHOUN, J. L. YANOWITZ, and P. D. SCHEDL, 1999  Novel functions of nanos in downregulating mitosis and transcription during the development of the Drosophila germline. Cell 99:271-281.[Medline]

DESHPANDE, G., L. SWANHART, P. CHIANG, and P. SCHEDL, 2001  Hedgehog signaling in germ cell migration. Cell 106:759-769.[Medline]

FISCHER-VIZE, J. A., G. M. RUBIN, and R. LEHMANN, 1992  The fat facets gene is required for Drosophila eye and embryo development. Development 116:985-1000.[Abstract]

FLEISCHMAN, R. A., 1993  From white spots to stem cells: the role of the Kit receptor in mammalian development. Trends Genet. 8:285-290.

FLYBASE,, 1999  The FlyBase database of the Drosophila genome projects and community literature. The FlyBase Consortium. Nucleic Acids Res. 27:85-88.

FOLEY, K. and L. COOLEY, 1998  Apoptosis in late stage Drosophila nurse cells does not require genes within the H99 deficiency. Development 125:1075-1082.[Abstract]

FORBES, A. and R. LEHMANN, 1998  Nanos and Pumilio have critical roles in the development and function of Drosophila germline stem cells. Development 125:679-690.[Abstract]

FUKUSHIMA, N., I. ISHII, J. J. CONTOS, J. A. WEINER, and J. CHUN, 2001  Lysophospholipid receptors. Annu. Rev. Pharmacol. Toxicol. 41:507-534.[Medline]

FULLILOVE, S. L., and A. G. JACOBSON, 1978 Embryonic development: descriptive, pp. 105–227 in The Genetics and Biology of Drosophila, edited by M. ASHBURNER and T. R. F. WRIGHT. Academic Press, New York.

GOETZL, E. J. and S. AN, 1998  Diversity of cellular receptors and functions for the lysophospholipid growth factors lysophosphatidic acid and sphingosine 1-phosphate. FASEB J. 12:1589-1598.[Abstract/Free Full Text]

GOETZL, E. J., H. LEE, H. DOLEZALOVA, K. R. KALLI, and C. A. CONOVER et al., 2000  Mechanisms of lysolipid phosphate effects on cellular survival and proliferation. Ann. NY Acad. Sci. 905:177-187.[Medline]

HAY, B., L. ACKERMAN, S. BARBEL, L. Y. JAN, and Y. N. JAN, 1988  Identification of a component of Drosophila polar granules. Development 103:625-640.[Abstract/Free Full Text]

HOLDER, N. and R. KLEIN, 1999  Eph receptors and ephrins: effectors of morphogenesis. Development 126:2033-2044.[Abstract]

HOLMES, A. L., R. N. RAPER, and J. S. HEILIG, 1998  Genetic analysis of Drosophila larval optic nerve development. Genetics 148:1189-1201.[Abstract/Free Full Text]

JACOBSON, M. D., M. WEIL, and M. C. RAFF, 1997  Programmed cell death in animal development. Cell 88:347-354.[Medline]

JAGLARZ, M. K. and K. R. HOWARD, 1994  Primordial germ cell migration in Drosophila melanogaster is controlled by somatic tissue. Development 120