The Posterior Determinant Gene nanos Is Required for the Maintenance of the Adult Germline Stem Cells During Drosophila Oogenesis

Corresponding author: Krishna Moorthi Bhat, Department of Cell Biology, Emory University School of Medicine, Atlanta, GA 30322., kbhat{at}cellbio.emory.edu (E-mail)

Communicating editor: R. S. HAWLEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In a variety of tissues in eukaryotes, multipotential stem cells are responsible for maintaining a germinal population and generating a differentiated progeny. The Drosophila germline is one such tissue where a continuous supply of eggs or sperm relies on the normal functioning of stem cells. Recent studies have implicated a possible role for the posterior determinant gene nanos (nos) in stem cells. Here, I report that nanos is required in the Drosophila female germline as well as in the male germline. In the female, nos is required for the functioning of stem cells. In nos mutants, while the stem cells are specified, these cells divide only a few times at the most and then degenerate. The loss of germline stem cells in nos mutant mothers appears to be due to a progressive degeneration of the plasma membrane. Furthermore, following germ cell loss, the germaria in the nos mutant mothers appear to carry on massive mitochondrial biogenesis activity. Thus, the syncytia of such germaria are filled with mitochondria. In the male germline, the male fertility assay indicates that nos appears to be also required for the maintenance of stem cells. In these mutant males, spermatogenesis is progressively affected and these males eventually become sterile. These results indicate novel requirements for nos in the Drosophila germline.

IN multicellular eukaryotes, several lineages, such as the haematopoetic lineage, skin, and intestinal epithelia, have a special type of cells known as stem cells (see DEXTER and SPOONCER 1987 ; HALL and WATT 1989 ). In these lineages, the pluripotential stem cells undergo asymmetric mitosis, self-renewing and, at the same time, producing another cell that is committed to a differentiation pathway. The germline of the fruitfly Drosophila is one such lineage (BROWN and KING 1964 ; KOCH and KING 1966 ; WIESCHAUS and SZABAD 1979 ; LIN and SPRADLING 1993 ; BHAT and SCHEDL 1997A ). The stem cells in this lineage undergo asymmetric divisions giving rise to a daughter cell that is the self-renewing stem cell and another daughter cell, the cystoblast (KING 1970 ; MAHOWALD and KAMBYSELLIS 1980 ). In females, the cystoblast is committed to differentiate into an egg. It divides four times to form a cluster of 16 interconnected cells called cystocytes. One of these cystocytes assumes the oocyte fate while the others become nurse cells and support the growth and differentiation of the oocyte.

The formation of pole cells, the precursor cells of the germline, during nuclear cycle 8 of embryogenesis marks the first visible step in germline development in Drosophila (see KING 1970 ; MAHOWALD and KAMBYSELLIS 1980 ). During gastrulation, these cells are carried along the ventral furrow and then through the invaginating gut to ultimately reach the presumptive gonadal site. By stage 13 of embryogenesis, these large germ cells, together with the somatic mesodermal cells, assemble into primitive gonads (see MAHOWALD and KAMBYSELLIS 1980 ). In females, although germ cells (prestem cells) divide two to three times before the late third instar larval stage, these divisions are simple mitoses, indicating that these germ cells have not yet acquired a stem cell identity. In males, on the other hand, the primitive gonads in the newly hatched first instar larvae already contain primary spermatocytes, although it is not clear whether these primary spermatocytes are generated from a stem cell division (i.e., asymmetric division) or from direct differentiation of a subset of germ cells (see BHAT and SCHEDL 1997A ). In both males and females, the primitive gonad initiates its overt differentiation during the transition of larvae to pupae. In females, this differentiation is much more elaborate than in males. First, the somatic cells of the primitive gonad in females begin to organize the gonad into 16–17 egg-producing units called ovarioles, with each ovariole sequestering five to seven prestem cells (see KING 1970 ). Second, a subset of the prestem cells assumes stem cell identity (i.e., ability to divide asymmetrically) while the remaining prestem cells differentiate directly into eggs without going through a stem cell stage (KING 1970 ; BHAT and SCHEDL 1997A ).

Despite many studies on Drosophila germline development, very little is known about how germline stem cells are formed or how they function. For instance, although a large number of female sterile mutations have been isolated (cf., SCHUPBACH and WIESCHAUS 1989 , SCHUPBACH and WIESCHAUS 1991 ), none of these mutations specifically disrupt the development of stem cells. While it has been recently shown that mutations in pumilio and piwi affect stem cell maintenance (LIN and SPRADLING 1997 ), the underlying reasons for the stem cell defect observed in these mutations remain unexplored. While the previous studies (BROWN and KING 1964 ; WIESCHAUS and SZABAD 1979 ; BHAT and SCHEDL 1997A ) indicate that each ovariole typically contains between one to three stem cells, electron microscopic (EM) studies (see KING 1970 ) suggest that each ovariole initially acquires at least four and perhaps as many as seven prestem cells. The extra prestem cells somehow appear to bypass the normal stem cell–cystoblast pathway for egg formation and, instead, differentiate directly into eggs (see KING 1970 ; BHAT and SCHEDL 1997A ). Therefore, the first few eggs deposited by a female are expected to be from prestem cells. Thus, in a "stem cell-specific" mutation, it is reasonable to assume that a few eggs will be produced from prestem cells.

It was recently shown that strong loss-of-function alleles of the posterior determinant gene nanos (nos) produced only a few eggs and that this defect can be rescued by a nos transgene (GAVIS and LEHMANN 1992 ; WANG et al. 1994 ; CURTIS et al. 1995 ). These authors suggested that nos affects the functioning of either the stem cells or the cystoblasts in the female germline (LEHMANN and NUSSLEIN-VOLHARD 1991 ; WANG et al. 1994 ; FORBES and LEHMANN 1998 ). nos was initially identified as a maternal effect mutation that affects the development of the abdominal region (TAUTZ 1988 ; STRUHL 1989 ; LEHMANN and NUSSLEIN-VOLHARD 1991 ; WANG and LEHMANN 1991 ). During embryogenesis, the Nos protein is concentrated at the highest level in the posterior pole plasm (see WANG et al. 1994 ). This posteriorly localized Nos protein promotes abdominal development by repressing translation of the maternally deposited hunchback (hb) mRNA (HULSKAMP et al. 1989 ; IRISH et al. 1989 ; STRUHL 1989 ). Hb protein is a morphogen and normally represses expression of abdominal gap genes (HULSKAMP et al. 1990 ; STRUHL et al. 1992 ). In the posterior region, a lower level of Hb (due to translational repression by Nos) allows normal development of the abdomen. In addition to its role in abdominal development, nos appears to be also required for the organization of pole cells into primitive gonads during late embryogenesis (KOBAYASHI et al. 1996 ). Consistent with a role for nos as a translational repressor, pole cells derived from embryos of mothers homozygous for one of the mutant alleles of nos, nos^BN, were found to express marker genes prematurely, and these germ cells failed to make a primitive embryonic gonad (KOBAYASHI et al. 1996 ). However, the reason behind the egg-laying defect in some of the stronger alleles of nos mutants was not known.

In this study, the underlying cause for the egg-laying defect displayed by females homozygous for the loss-of-function mutations in nos has been investigated. The egg-laying defect in nos mothers appears to be due to the degeneration of stem cells in the adult germline. In the absence of nos activity, the identity of germline stem cells is specified and the cells undergo a few asymmetric divisions. These stem cells then degenerate. The stem cell death in nos mutants is caused by an age-dependent, progressive loss of the plasma membrane. The results also show that following germ cell loss, the germarium promotes a massive mitochondrial biogenesis activity. Finally, our results indicate that loss of nos activity in males affects spermatogenesis and these males progressively become sterile. These findings uncover novel requirements for nos during germline development in Drosophila.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks, genetics:
The two ethyl-methyl-sulfonyl fluoride-induced nos alleles, nos^RD and nos^RC, and a deficiency for nos gene {Df (3R) Dl^Fx3}, were obtained from Dr. Ruth Lehmann (LEHMANN and NUSSLEIN-VOLHARD 1991 ). nos^RD is a strong partial loss-of-function allele and nos^RC appears to be a null allele (see WANG et al. 1994 ). cyclin A mutant and a deletion mutant for cyclin B were obtained from the Drosophila stock center. The flies were raised at room temperature on standard corn fly food. All the experiments were done at room temperature.

Egg-laying pattern analysis:
Freshly eclosed females (wild type, nos homozygotes, and transheterozygotes for different nos mutant alleles) were mated singly with 2-day-old virgin wild-type males. These were transferred to new vials every 24-hr, and the number of eggs laid within this period were counted for ~4 wk. Mutant females were classified as Type I, II, or III and the mean number of eggs laid per day by each type was plotted against days. For each allelic combination (nos^RD/nos^RD, nos^RD/nos^RC, nos^RD/Df (3R) Dl^Fx3, and nos^RC/Df (3R) Dl^Fx3), 18 separate matings were examined (the total number of females examined, N = 72). The data presented represent the egg-laying patterns of mothers transheterozygous for nos^RD and the deficiency for nos, Df (3R) Dl^Fx3. However, various nos allelic combinations showed all three types of egg-laying patterns, though the percentage of mutant females in each type differed depending on the allelic combination.

Male fertility assay:
Freshly eclosed nos mutant males {nos^RD/Df (3R) Dl^Fx3 and nos^RC/Df (3R) Dl^Fx3} were individually challenged with wild-type virgin females every day for 35 days. Wild-type males were similarly challenged with virgin wild-type females as controls. The fertility of each male >35 days was determined by the presence/absence of progeny in each of the vials (see also legend to Table 1). A minimum of 30 males or females of each genotype was tested. In these male fertility assays, mutant males were often able to fertilize eggs; however, these eggs did not develop. These time points were counted as "positive" in the assay, whereas those time points at which the females were not inseminated by a mutant male (indicated by a lack of egg deposition by a female) were considered "negative."

Estimation of the primary spermatocyte number:
Testis from nos^RD/Df (3R) Dl^Fx3, nos^RC/Df (3R) Dl^Fx3, and Df (3R) Dl^Fx3/+ (as control) were dissected, fixed in 4% paraformaldehyde, and permeabilized in 75% glycerol overnight. The primary spermatocytes with their characteristic prominent nucleolus were counted in an Axioplan microscope at high magnification (x1000). In another approach, several single level photomicrographs were taken and the cells showing a prominent nucleolus were counted. This approach was adopted to avoid the possibility of counting the same primary spermatocyte more than once. The number obtained from this counting is a very conservative quantification of the primary spermatocytes in wild-type or mutant testis and are the ones given in the text. About threefold higher numbers were obtained by counting using a microscope; however, the difference between the various allelic combinations and the wild type remained more or less the same in either of the two ways of counting. The numbers are estimated from 6 to 10 testes. Also, the extreme cases where most of the testes were filled with primary spermatocytes (see Figure 3B) were not used in this estimation.

View larger version (93K): In this window In a new window Download PPT slide   Figure 1. Ovarian phenotypes in nanos mutants. (A) Ovaries from wild type and (B–F) nosRD/nosDf mutant mothers were dissected, stained with the DNA stain Hoechst, and examined under a UV fluorescent microscope. g, germarium; ec, egg chamber; DA, dorsal appendage. Except in F, which represents an ovary from an ~7-day-old mutant mother, ovaries are from mutant mothers aged between 2 and 3 days.

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. nanos mutants show three different types of disrupted egg-laying patterns. In Type I, nos mutant females begin to deposit their eggs by the second day of emergence, as in wild type. They stop laying eggs usually by the fifth day. In Type II, nos mutants begin to deposit eggs by the second day of emergence and cease generally around the fifth day; after an interval of 2 to 12 days of no egg deposition, Type II females lay a few more eggs. In Type III, nos mutant females lay no eggs at all. Because not all ovarioles produce egg chambers in a given nos female, the number of eggs laid by Type I and Type II females are fewer than those laid by wild-type females.

View larger version (108K): In this window In a new window Download PPT slide   Figure 3. nanos is also required for the male germline development. Fixed testes were examined by Nomarski optics. Anterior end is toward the left; S, sperm bundles. The anterior region containing mostly the primary spermatocytes is marked by lines. A, C, and D are of the same magnification. Stem cells occupy the very tip of the testis. (A) A testis from a wild-type male. (B and C) Testes from nosRC/nosDf males. Note that the entire testis in B is filled with primary spermatocytes. (D) A testis from a nosRD/nosDf male.

Immunohistochemistry:
Groups of 10–12 freshly eclosed virgin females were mated and kept in vials for different durations; their ovaries were dissected, fixed in 4% paraformaldehyde for 25 min, and stained with anti-Sxl antibody (BOPP et al. 1993 ) as described previously (BHAT et al. 1996 ; BHAT and SCHEDL 1997A , BHAT and SCHEDL 1997B ) and examined by a confocal microscope. Because each ovary is made up of 16–17 egg-producing units called ovarioles, to determine the frequency of each mutant phenotype, the ovarioles are separated from one another and each ovariole is examined by a confocal microscope. The total numbers of ovarioles examined were as follows: 320 from 0.5-day-old females; 280 from 1 day; 270 from 2 day; 170 from 3 day; 290 from 4 day; 190 from 11 day. The mutant ovaries were also examined for the expression of several other markers: Orb (LANTZ et al. 1994 ), Vasa (LASKO and ASHBURNER 1990 ), GAGA/Trithorax-like (BHAT et al. 1996 ), bag-of-marbles (bam), and a cystoblast specific marker (MCKEARIN and SPRADLING 1990 ), Cyclin B and ß-Tub.

Hoechst staining:
For Hoechst staining, ovaries and testes were fixed in 4% paraformaldehyde for 20 min, washed in PBS, and stained in a 1-µg/ml solution of Hoechst for 10 min. They were then washed several times with PBS, transferred to a 50% glycerol:PBS solution for 2–3 hr, and kept in a 75% glycerol:PBS solution overnight at 4°. The samples were then mounted on a slide between coverslips and examined under UV light. For Nomarski Optics photomicrography, ovaries and testes were fixed in 4% paraformaldehyde, washed in PBS, and mounted in a 75% glycerol:PBS solution.

Electron microscopy:
Ovaries were dissected in 0.1 M Na cacodylate buffer, fixed in Karnofsky's fixative (3% gluteraldehyde, 1% paraformaldehyde in 0.1 M Na cacodylate) with 0.05% saponin for 2 hr at room temperature. Ovaries were then washed in 0.1 M Na cacodylate buffer, fixed again in 2% osmium tetroxide with 0.05% K Fe Cn in 0.1 M Na cacodylate for 1 hr at room temperature. Next ovaries were washed with 0.05 N maleate buffer, prestained in 0.5% uranyl acetate in 0.05 N maleate buffer at 4° overnight, then rinsed in distilled water, dehydrated, and infiltrated first in resin (Quetol/Epoxy 812) and acetone mixture in a stepwise manner—1/1 for 2 hr, 1/2 for 2 hr, 1/3 overnight on dri-rite, and then in resin alone for 2 hr. Embedding was done in plastic molds at 45° overnight and then at 75° overnight. These were then sectioned and examined under an electron microscope.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The nanos mutant alleles:
Previous results indicate that females homozygous for two of the strong loss-of-function mutant alleles of the posterior determinant genes nos, nos^RC, and nos^RD (henceforth called nos mutant females) produced only a few eggs (LEHMANN and NUSSLEIN-VOLHARD 1991 ) and that this defect can be rescued by a nos transgene (GAVIS and LEHMANN 1992 ; WANG et al. 1994 ; CURTIS et al. 1995 ). In nos^RC, Nos protein was undetectable in the germarium, nurse cells, or in embryos (WANG et al. 1994 ) and could be a null allele of nos (FORBES and LEHMANN 1998 ). In nos^RD, the amount of Nos protein was reduced in the germarium and in the embryo, whereas the level was normal in stage-10 nurse cells (see WANG et al. 1994 ). These alleles appear to behave genetically as loss-of-function alleles of nos; however, nos^RC appears to be a much stronger allele than nos^RD (see below).

Ovarian phenotypes in nanos mutant mothers:
To determine the ovarian phenotypes in nos mutant females, initially ovaries from nos mutant mothers were stained with Hoechst, a DNA stain, and examined under a fluorescent microscope. In wild type, each ovariole contains six to seven developing egg chambers of different stages and generally one, but occasionally two, mature eggs (Figure 1A). In nos mutant mothers, as shown in Figure 1B, Figure C, and Figure D, the ovaries contained ovarioles with either no egg chambers or a single (Figure 1, C–F) to a maximum of seven egg chambers of different stages (Figure 1B). The mutant ovarioles shown in Figure 1D–F, indicate that some of the ovarioles in nos mutants produce only one egg. The Hoechst staining of mutant ovaries also indicated that ovaries with no developing germ cells can also be observed among nos mutant females (data not shown). The ovarian phenotypes described here were observed in several different nos allelic combinations (nos^RD/nos^Df, nos^RC/nos^Df, and nos^RD/nos^RC). Thus, these results are consistent with the previous observation that ovarioles in nos ovaries produce a variable number of eggs (CURTIS et al. 1995 ) and argue that the egg-laying defect in nos mothers is due to an interruption in the oogenesis process itself.

nanos mutant females exhibit three different types of interrupted egg-producing patterns:
Because the strong loss-of-function alleles of nos produced only a few eggs, it is possible that the germline stem cells are not specified in nos mutants and that the few eggs deposited by the mutant females were due to a subset of prestem cells that differentiated directly without going through a stem cell stage (see KING 1970 ; BHAT and SCHEDL 1997A ). Alternatively, one can also suppose that the loss of nos function affects the functioning but not the specification of stem cells and that the few eggs generated were due to both direct differentiation of prestem cells and to stem cells asymmetrically dividing a few times to produce a limited number of eggs.

To distinguish between these two possibilities, first the egg-laying pattern of nos mutant females was determined. Because each asymmetric division of a stem cell yields one egg, determining the pattern and the number of eggs deposited over a period of time in a given mutation would indicate the history of functioning of stem cells in that mutation. In wild type, a stem cell divides once in 12–22 hr (see MAHOWALD and KAMBYSELLIS 1980 ) to self-renew and to produce a cystoblast. A cystoblast undergoes 6–7 days of development before being deposited as an egg. A female Drosophila begins to lay eggs by its second day of emergence, indicating that parent cystoblasts for these eggs enter the differentiation pathway in the early pupal stage (the pupal stage extends for ~5 days at room temperature). Generally, a wild-type female deposits an average of 30–40 eggs per day for ~4 wk, after which the number becomes progressively reduced. As shown in Figure 2, nos females showed three different types of disrupted egg-producing patterns. In Type I, nos females began to deposit their eggs by the second day of emergence as in wild type, indicating that the process of egg development was not delayed in the mutants. However, these nos females stopped laying eggs altogether by their fifth to ninth day of emergence. In Type II, nos mothers began to deposit their eggs by their second day of emergence and ceased to lay any eggs between the fifth to ninth day, as in Type I; however, after an interval of 2 to 12 days of no egg deposition, these females laid a few more eggs (Figure 2). In Type III, nos mutant females did not lay any eggs. Phenotypic analysis of these mutant ovaries indicated that the ovarioles in Type III individuals are rudimentary and devoid of any developing egg chambers or other developing germ cells.

As shown in Figure 1C–F, the phenotypic examination of nos ovaries also indicated that not all ovarioles in a given nos mutant female produce egg chambers. This appears to account for the fewer eggs laid by Type I and Type II females. Finally, as shown in Table 1, the percentage of mutant females showing the three different types of egg-laying patterns between the two nos mutant alleles and a deficiency that uncovers the nos gene (nos^RC/nos^Df vs. nos^RD/nos^Df) indicated that nos^RC is likely to be a stronger loss-of-function allele of nos (see also WANG et al. 1994 ). For instance, a majority of mutant females of genotype nos^RD/nos^Df showed a Type II egg-laying pattern, whereas nearly equal percentages of mutant females of genotype nos^RC/nos^Df showed Type I and Type II egg-laying patterns. From these results, it seems likely that nos^RC/nos^Df represents either a complete loss-of-function, or at the minimum, a strong loss-of-function allelic combination for nos.

nanos is also required in the male germline:
To determine whether nos plays a role in the male germline development, testis from nos mutant males were dissected, fixed, and examined by Nomarski optics. When nos^RC/nos^Df males were examined, nearly 20% of the mutants showed morphologically distinct testis defects. As shown in Figure 3B and Figure C, the testis from these males contained fewer sperm bundles and, at the same time, had a large pool of primary spermatocytes (with their characteristic prominent nucleolus), often filling most of the anterior region of the testis. A very conservative counting of the number of primary spermatocytes (see MATERIALS AND METHODS) in the mutant testis indicated that the number of primary spermatocytes in the mutant testis from nos^RC/nos^Df males was 157 (the number of defective testis examined, N = 7), whereas in wild type the number of primary spermatocytes was only 45 (N = 6). In nos^RD/nos^Df males, the number of primary spermatocytes was also enhanced and found to be ~108 (N = 6). Rarely, nos^RC/nos^Df testes with no sperm bundles were observed and such testes were entirely filled with primary spermatocytes (Figure 3B). Moreover, a defective testis in nos mutant males generally had fewer coils compared to wild type, a likely consequence of the presence of fewer sperm bundles within the testes. Similar defects were also observed in males of genotype nos^RD/nos^Df (Figure 3D); however, the defects were much less severe. These defects were also observed in males of nos^RD/nos^RC (data not shown).

We also examined spermatogenesis in nos mutant males by determining the ability of these males to successfully inseminate females in a male fertility assay over a period of ~4 wk (see MATERIALS AND METHODS). For this experiment, we used the nos^RC/nos^Df males. As shown in Table 1, ~5% of the males (N = 30) were unable to inseminate females throughout the length of time they were tested. These males, therefore, were classified as Type III (see Table 1). However, an important difference must be noted between Type III females and Type III males. In Type III mutant females, all of the ovarioles are devoid of any germ cells, whereas in Type III males, the testis always had germ cells, but they were either undifferentiated primary spermatocytes (Figure 3B) or the sperm that were present were presumably defective. About 90% of the males showed a Type II defect, whereas 5% showed a Type I defect. In all these cases, the male fertility was reasonably high for the first 3 days; however, it dropped significantly after that. All of the males examined became sterile by the third week. In summary, these results indicate that nos is required in the male germline as well.

The age-dependent progression of germline defects in nanos mutant ovarioles:
To determine the cause for the cessation of egg production in the nos mutant mothers, we next examined the mutant ovaries with several cell-type specific markers. Since the egg-laying patterns in nos mutant mothers were influenced by the age of the individuals (see Figure 2), ovaries from mutant mothers that were aged for different durations were stained with various antibodies. Figure 4, for example, shows nos mutant ovaries stained with an antibody against the Sex lethal (Sxl) protein. The stem cells in an adult ovariole are located in the anterior-most region within the germarium (see KING 1970 ; LIN and SPRADLING 1993 ). In wild type, the Sxl protein is present at very high levels in the cytoplasm of stem cells (Figure 4A, sc). It is also present in the cytoplasm of cystoblasts (cb) and two-cell cysts. The Sxl expression is downregulated when two-cell cystocytes become four-cell cystocytes and are no longer detectable in the cytoplasm (see BOPP et al. 1993 ). When ovaries from 0.5-day-old mutant mothers were stained with Sxl, nearly 40% of the ovarioles (see Table 2; Figure 4B) were indistinguishable from wild-type ovarioles. These mutant ovarioles contained stem cells, cystoblasts, and cystocytes within the germarium. Their vitellaria contained zero to a few egg chambers. Furthermore, the germaria of these ovarioles were normal in size, which is an indication that they contain several clusters of developing cystocytes (older than the two-cell stage) and stage-1 egg chambers inside. These results indicate that many ovarioles in nos mutant mothers start out as normal ovarioles.

View larger version (70K): In this window In a new window Download PPT slide   Figure 4. Age-dependent progression of germline defects in nanos mutants. The ovaries are stained with Sxl antibody. sc, stem cells; cb, cystoblasts; ec, egg chambers; DA, dorsal appendage; g, germarium. The germarium is divided into four regions: R1, R2a, R2b, and R3. Stem cells occupy the anterior-most position within the germarium in R1. R1 and R2a contain cystoblasts and 2- to 8-cell cystocytes. R2a–R3 contain several 16-cell cystocytes. R3 also contains a stage-1 egg chamber. (A) Stem cells (located at the anterior-most region of the germarium), cystoblasts, and 2-cell cysts expressing high levels of Sxl are shown. (B) Shown is a wild-type-looking mutant ovariole from a 0.5-day-old mutant female. (C) Mutant ovarioles from a 0.5-day-old female with either 1–3 stem cells or none; these ovarioles also lacked egg chambers in the vitellarium (see text). (D) An ovariole from a 4-day-old nos mutant female; the germarium of this ovariole has no Sxl-positive cells, whereas the vitellarium has few egg chambers. (E and F) Ovarioles from 4-day-old nos mutant females. The germarium in these ovarioles has (E) one and (F) three stem cells, respectively. (G and H) Ovarioles from 11-day-old mutant females. (G) Rarely, an ovariole with a stage-14 egg chamber (only the dorsal appendage is visible) attached to the empty germarium can be observed. (H) Nearly all of the ovarioles from 11-day-old mutant females had empty germarium with no germ cells inside and no egg chambers in the vitellarium. See Table 2 and text for frequencies of occurrence of these phenotypes.

In ~10% of the ovarioles in ovaries from 0.5-day-old nos mothers, the germarium contained one to three stem cells but no cystoblasts or two-cell cystocytes (Table 2; Figure 4C). That these large cells at the tip of the germarium are indeed stem cells and not cystoblasts is indicated by the lack of expression of bag of marbles (bam) in a whole-mount RNA in situ experiment (data not shown; see DISCUSSION). bam is expressed in the cystoblast but not in the stem cells of the female germline, and thus serves as a cystoblast marker (MCKEARIN and SPRADLING 1990 ). The absence of cystoblasts and cystocytes in these germaria would argue that stem cells have not divided to produce any cystoblasts and that nos may be required for stem cell division. In another 10% of the ovarioles, the rudimentary germarium had no Sxl-positive stem cells, cystoblasts, or two-cell cysts (see Table 2; Figure 4D). However, the vitellarium of these ovarioles contained several developing egg chambers (Figure 4D), which are likely to have originated from the direct differentiation of prestem cells (BHAT and SCHEDL 1997A ). The remaining 40% of the ovarioles had neither germ cells inside their rudimentary germarium nor any developing egg chambers in the vitellarium (Figure 4D; Table 2), indicating that no germ cells underwent development in these ovarioles.

The mutant ovaries from 2-day-old nos mothers were next examined with anti-Sxl (BOPP et al. 1993 ). As shown in Table 2, the normal-looking ovarioles in these ovaries had decreased by 10% to 30% and empty ovarioles had increased from 10% to 50%. The germarium of ~5% of the ovarioles had only stem cells in region 1 (see Table 2; cf., Figure 4C). Germ cells that would account for cystoblasts or two-cell cystocytes were not observed (data not shown), indicating that these stem cells had failed to divide to generate any cystoblasts. In another 5% of the ovarioles, the germarium had only stem cells; however, the vitellarium had one or two egg chambers (see Table 2).

When ovaries from 5-day-old nos mothers were examined (see Table 2), a striking change in the frequency of occurrence of various types of ovarioles was observed. For instance, the percentage of normal-looking ovarioles decreased from ~40% in 0.5-day-old mutants to 5%. Consequently, the ovarioles that lacked stem cells, cystoblasts, and cystocytes in the germarium but contained developing egg chambers in the vitellarium (cf., Figure 4D) increased to 35%. This indicates that the germarium in a large number of ovarioles became nonfunctional by the fifth day, which can be attributed to loss of stem cells. Furthermore, as shown in Table 2, the percentage of ovarioles with only stem cells in the germarium but a few developing egg chambers in the vitellarium (presumably originating from the cystoblasts that were generated much earlier) had marginally increased from 8% in 3-day-old females to 10%. In such ovarioles, the number of stem cells at the tip of the germarium was found to range between one to three (Figure 4E and Figure F), although two stem cells were more frequent than one or three. The absence of cystoblasts or developing cystocyte clusters in these rudimentary germaria indicates that these stem cells are not dividing. Moreover, stem cells in many of the ovarioles in these 5-day-old mutant mothers had very weak Sxl expression (Table 2).

When ovaries from 11-day-old nos mothers were stained with Sxl, the percentage of wild-type-looking ovarioles had decreased to zero, and nearly 93% of the ovarioles did not contain any germ cells in the germarium or egg chambers in the vitellarium (Table 2; Figure 4H). A small percentage of ovarioles had only egg chambers in the vitellarium (Table 2; Figure 4G). These results indicated that the germaria eventually lost stem cells. Because no new cystoblasts had been generated in these ovarioles, these ovarioles had become agametic.

A similar pattern of progression of these germline defects was observed when ovaries from mutant females aged for different durations were stained with vasa, another germ cell marker (LASKO and ASHBURNER 1990 ), Trithorax-like/GAGA, a chromatin protein (BHAT et al. 1996 ), or propidium iodide, a DNA stain (data not shown). The above results are also consistent with the conclusion drawn based on the Type II egg-laying pattern that loss of nos activity affects the functioning of the germline stem cells. Thus, we conclude that the stem cells are indeed formed in nos mutants; however, they either fail to divide or divide only transiently, leading to a cessation of egg production (see below).

The germline stem cells function only transiently in nanos mutants:
The results described thus far indicate that in the absence of nos activity, a subset of the germ cells assumes stem cell identity; however, these stem cells function only transiently. To obtain additional evidence to support this possibility, nos ovarioles were examined with ovarian RNA-binding protein (Orb) antibody (see LANTZ et al. 1992 ). In wild type, while the Orb protein can be detected at low levels in the cytoplasm of stem cells and cystoblasts (cf., Figure 5B), it subsequently localizes to the oocyte within a 16-cell cystocyte cluster (Figure 5A). Thus orb is a reliable marker for identifying postcystoblast germ cell development within the germarium.

View larger version (60K): In this window In a new window Download PPT slide   Figure 5. Stem cells divide only a few times in nanos mutants. The ovaries are stained with Orb antibody. sc, stem cells; cc, cystocytes; ec, egg chambers; DA, dorsal appendage; g, germarium. (A) An ovariole from a 3-day-old wild-type female with the germarium packed with developing cystocytes. The stem cells stain very weakly with anti-Orb; however, in A, they are out of the focal plane. (B) Ovariole from a 3-day-old nos mutant female. Note the empty region (open arrow) between stem cells (at the tip of the germarium) and a cystocyte cluster and the region between this cystocyte cluster and the stage-1 egg chamber in region 1. A few egg chambers can be observed in the vitellarium of such ovarioles. (C and D) Shown are ovarioles from 5-day-old mutant females. Note the empty region (open arrow) between a cystocyte cluster and a stage-1 egg chamber in C. Stem cells and cystoblasts are no longer detectable. In D, only a stage-1 egg chamber (in R3) is present; no cystocytic clusters (in R2-3) or stem cells (in R1) are observed. (E) An ovariole from a 7-day-old nos mutant female is shown. The germarium is empty; however, the vitellarium contains a few egg chambers. (F) An ovariole from an 11-day-old nos mutant female. A stage-14 egg chamber (only the dorsal appendage is visible) attached to the empty germarium is observed.

In wild type, stem cells regularly divide and produce cystoblasts as cystocytes move down the germarium. Therefore, the germarium is tightly packed with developing germ cells, as can be visualized with Orb staining (see Figure 5A). When ovarioles from 2- to 3-day-old mutant females were examined with Orb, germaria with regions that were devoid of developing germ cells were observed (Figure 5B and Figure C, open arrows). As shown in Figure 5B, while two to three stem cells can be seen at the tip of the germarium, the region immediately posterior to these stem cells had no cystoblasts and cystocytes. This staining pattern would suggest that stem cells in this germarium had not divided to generate cystoblasts. Furthermore, in a mutant germarium, cystocyte clusters in regions 2–3 and/or a stage 1 egg chamber in region 3 and a few older-stage egg chambers in the vitellarium were frequently observed (see Figure 5D). Also, mutant ovarioles with their completely agametic germaria and vitellaria with a few developing egg chambers were also observed (Figure 5E and Figure F). The presence of egg chambers in these ovarioles indicates that stem cells earlier produced a few cystoblasts in these ovarioles. Moreover, a pattern of interrupted cystoblast production in the mutant germaria can be discerned on the basis of the presence of agametic regions between a cystocyte and a stage-1 egg chamber (open arrows in Figure 5B and Figure C). Because germaria with intervening agametic "regions" were not observed in the older nos females (Figure 5, D–F), it appears that the stem cells divide only a few times early on in their life span to generate a few cystoblasts. These interpretations are consistent with the result that nos mothers exhibit an interrupted pattern of egg laying (see Figure 2).

The apparent slow rate of cystoblast production by stem cells (see also Figure 2) suggests that the division of stem cells in nos mutant mothers might be delayed. Thus, it is possible that nos affects the accumulation of gene products critical for entry into mitosis, such as Cyclin B. This is based on the observation that the Cyclin B message contains a nos-response element (NRE) in the 3'-untranslated region (see DALBY and GLOVER 1993 ) similar to those observed in hb and bicoid. Nos is thought to repress the translation of hb and bicoid messages via NRE. If Nos regulates Cyclin B translation in the germline stem cells, the expectation would be that Cyclin B levels will be downregulated in stem cells. However, no significant differences in the levels of Cyclin B between nos mutant stem cells and wild type were observed (Figure 6). Moreover, nos mutants did not show any genetic interaction with cyclin B mutants (nos/nos;cyclin B/+) in the germline. These results are consistent with the previous finding by DALBY and GLOVER 1993 that injection of truncated cyclin B mRNAs without the 3'-untranslated region to wild-type embryos caused translational derepression of the injected mRNAs; however, injection of wild-type cyclin mRNAs to nos mutant embryos did not result in a derepression of the translation of the tagged cyclin mRNAs.

View larger version (30K): In this window In a new window Download PPT slide   Figure 6. The levels of Cyclin B protein in stem cells are not affected in nanos mutant females. The ovaries are stained with an antibody against Cyclin B (red) and the DNA dye, YOYO (green), to visualize the nucleus. g, germarium. (A and B) Merged images of the same confocal section from (A) a wild-type ovariole and (B) an ovariole from a nos mutant.

Germ cells suffer from a membrane defect in nanos mutants:
The results described thus far indicate that germline stem cells in nos mutant ovaries divide only a few times at most during the life span of the mutant mother. However, it is not clear why stem cells cease to divide in nos mutants; nor is the ultimate fate of stem cells in these ovaries clear. One possibility is that stem cells undergo cell death in nos mutants. Therefore, germaria from newly eclosed (0.5-day-old) to 7-day-old nos mutant females were examined by EM. The EM studies indicate that, while the stem cells (and other developing germ cells) in the mutant germaria in 0.5-day-old females were wild type-looking, the stem cells in the germaria of 2- to 3-day-old-mutant females were not. The EM photographs of 2 to 3 day-old-mutant germarium revealed a striking plasma membrane defect in germ cells (Figure 7). As shown in Figure 7C–E, small localized lesions were present on the plasma membrane of stem cells. These lesions were infiltrated with what appears to be disintegrating membranous material (Figure 7E, arrowheads).

View larger version (202K): In this window In a new window Download PPT slide   Figure 7. Loss of nanos activity causes degeneration of the plasma membrane. Electron microscopy of wild-type and nos mutant ovarioles. Only the germarial sections are shown. sc, stem cells; cb, cystoblast; tf, terminal filament; N, nucleus; small arrows, plasma membrane; large arrows, nuclear membrane; long arrows, mitochondria; arrowheads, degenerating plasma membrane. (A) Wild-type germanium from a 4-day-old female (2.6 x 1000 magnification) showing stem cells, cystoblasts, and developing cystocyte clusters packed tightly inside. (B) A mutant germarium from a 4-day-old nos mutant female (3.3 x 1000). A presumably single stem cell is present at the tip of the germarium adjacent to the terminal filament. (C) Higher magnification (13 x 1000) of a portion of the germarium shown in B. Note that the plasma membrane of stem cells is degenerating in several places (white plaques on the plasma membranes indicated by arrowheads). (D) A mutant germarium (10 x 1000) showing initial stages of the plasma membrane degeneration (arrowheads). (E) In this mutant germarium (15 x 1000), the degeneration of the plasma membrane has progressed further. (F and G, 13 x 1000) Plasma membrane of cystoblasts and cyctocytes also degenerate in nos mutants (arrowheads).

The EM examination of the germaria from 2- to 3-day-old nos mutants also indicates that the plasma membrane of other germ cells, i.e., cystoblasts and cystocytes, also had similar plasma membrane defects. As shown in Figure 7F (region 2A of the germarium) and Figure 7G (region 2B of the germarium), the plasma membrane of cystocytes of various stages (2–16 cell stages) is disintegrating (arrowheads). However, the plasma membrane defect appears to be restricted to stem cells, cystoblasts, and cystocytes because such membrane defects were not observed in stage 1–14 egg chambers (data not shown).

To examine the progression of the plasma membrane defect of germ cells in nos mutants, germaria from 5- to 6-day-old mutant mothers were next examined by EM. As shown in Figure 8A–C, germaria with one or two nuclei at the tip often were observed in these ovaries. Since there was no trace of plasma membrane in these cases, we think that the plasma membrane of these cells had completely degenerated. Given the size of the nucleus and, more importantly, the location of the nuclei within the germarium, i.e., close to the tip, and also on the basis of our immunostaining results (see also Figure 4, Figure 5, Figure 6, and Figure 7C), we believe that these are stem cell nuclei. Remarkably, the nuclear membranes of these cells are intact (arrows in Figure 8, A–C). This indicates that in the nos mutant only the plasma membrane, but not the nuclear membrane, is affected. On the basis of these results, we suggest that the loss of stem cells in nos mutant females is due to disintegration of the plasma membrane and resulting cell degeneration.

View larger version (175K): In this window In a new window Download PPT slide   Figure 8. Loss of germ cells in nanos mutants causes a massive mitochondrial hyper-proliferation. N, nucleus; arrow, nuclear membrane; long arrow, mitochondria; tf, terminal filament. (A) A wild-type germanium from a 4-day-old female (2.6 x 1000 magnification). (B) A germarium from 6-day-old nos mutant females (2.3 x 1000). Only a nucleus and a large number of mitochondria can be observed in the syncytium (long arrow). (C) A mutant germarium shown at higher magnification (5 x 1000). (D) An ovariole from a 0.5-day-old Type III female.

The germaria in nanos mutants contain a large number of mitochondria:
The EM examination of germaria from 5-day-old nos females also reveals another remarkable phenotype. The germarium of such ovarioles contained an unusually large number of mitochondria (Figure 8, A–C, long arrows). Interestingly, this mitochondrial proliferation was not observed in ovarioles from younger nos mutant females prior to the loss of germ cells (Figure 7, B–E) but only observed in ovarioles of older nos females. To determine whether a prior presence of germ cells is required for mitochondrial hyper-proliferation in nos mutant germaria, rudimentary ovarioles from 0.5-day-old Type III nos females (ovarioles in Type III nos mutant mothers lack germ cells) were examined by EM. As shown in Figure 8D, mitochondrial hyper-proliferation was not observed in Type III nos ovarioles. Similarly, overproliferation of mitochondria was not observed in aged Type III mutant female ovaries (data not shown). A mitochondrial overproliferation was also not observed in ovaries from the daughters of mothers homozygous for the grandchildless mutant, tudor (data not shown). These results indicate that prior presence of germ cells is necessary for the overproliferation of mitochondria and that the existing mitochondria from these germ cells proliferate following the loss of germ cells within the germarium.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Stem cells, in general, have a few fundamental differences from other cell types. Besides being able to self-renew, stem cells undergo asymmetric mitosis and generate a progeny that is committed to a differentiation pathway. The results presented in this article provide some insights into the stem cells of the female germline in Drosophila. These results also uncover novel germline requirements for the posterior determinant gene nos. Of most significance is the finding that nos is required in stem cells of the adult female germline. In the Drosophila germline, previous studies indicate that nos is required for the migration of germ cells (KOBAYASHI et al. 1996 ). It is also reported to be required in the stem cells (FORBES and LEHMANN 1998 ). While we also independently came to the conclusion that nos is required in stem cells, this study extends the earlier results of FORBES and LEHMANN 1998 in several ways. Here, we show that nos is required for the functioning of stem cells and not essential for the formation of stem cells or stem cell identity specification. We find that nos is required for the survival of stem cells. In females lacking a wild-type nos activity, the germline stem cells undergo a few asymmetric divisions and then cease to divide. This study also shows that the cessation of stem cell division in nos mutants is due to the degeneration of these cells following a few asymmetric divisions. Moreover, this study addresses the underlying cause for the loss of stem cells: the degeneration of the plasma membrane. Finally, our results also indicate that nos is required in the male germline.

Nanos function in the male germline:
Our results indicate that when nos^RC/nos^Df males were subjected to a male fertility assay, most males showed fertility defects (Table 1). This indicates that nos is also required in the male germline. As in females, three distinct types of fertility defects, Type I, Type II, and Type III, can be discerned in nos mutant males. Consistent with these fertility defects, an examination of testis from the mutant males reveals testis defects (Figure 3) that range from having no sperm bundles to a few sperm bundles. Moreover, an abnormally high number of primary spermatocytes was observed in the anterior region of the testis, suggesting a delay or block in the differentiation of these cells (Figure 3). The large number of cells in the anterior region of the mutant testis (see Figure 3) are unlikely to be spermatids that had undergone an incomplete elongation. This conclusion is based on the fact that spermatids do not display a prominent nucleolus, unlike primary spermatocytes. It must also be noted that these defects were not observed in nos^Df/+ individuals.

By the phenotypic criterion, it seems that the effect of loss of nos on the male germline is somewhat different from the effect on the female germline and it is not at all clear whether nos affects male germline stem cells. That is, if the underlying germline defects were the same in both males and females, one would expect to observe empty testes devoid of any germ cells in nos mutants of advanced age. However, no such instances were observed. One of several possibilities for this difference could be due to male and female specific differences in the germline requirements for nos. Besides, the modes of development and functioning of the germline in males and females are quite different. For example, each cystoblast (or prestem cell) in the male germline produces 64 sperm, whereas in the female, only one egg is generated from a cystoblast. It therefore remains an open question whether nos is required in the male germline stem cells and what are the precise requirements of nos in poststem cell spermatogenesis steps.

The loss of nanos activity affects the survival but not the specification or the asymmetric division of stem cells in the germline:
The results presented in this article indicate that nos is required for the survival of stem cells in the female germline but not for their formation or specification. That stem cells are indeed specified in nos mutant ovaries is indicated by at least two different lines of evidence. First is the Type II egg-laying pattern displayed by nos mutant females (see Figure 2). While the first batch of eggs in Type II may also contain eggs originating from a direct differentiation of prestem cells (see BHAT and SCHEDL 1997A ), the second peak of eggs (which are laid as late as 12 days after eclosion; see Figure 2) are likely to have been generated from the asymmetric division of stem cells because all the prestem cells would have been differentiated and laid as eggs within the first 2–3 days of eclosion. Thus, the Type II egg-laying pattern indicates that stem cells are specified in nos mutant females and also that stem cells divide transiently to generate a few cystoblasts.

The second line of evidence comes from the staining pattern of the mutant germaria with several cell-type specific markers. In freshly eclosed females, the germaria in a large number of ovarioles in nos mutants appear to be wild type, with cells that would account for stem cells, cystoblasts, and two-cell cystocytes in region 1. In older females, however, only the stem cells can be observed at the tip of the germarium, not the cystoblasts or the cystocyte clusters (see Figure 4). That the large cells at the tip of the germarium are indeed stem cells is also indicated by their lack of bam mRNA (data not shown), a cystoblast specific marker (MCKEARIN and SPRADLING 1990 ). On the other hand, ovarioles of 4- to 5-day-old mutants containing only stem cells in the germarium and a few egg chambers in the vitellarium (Figure 4E and Figure F) indicate that stem cells have divided earlier to generate a few cystoblasts. Moreover, the presence of empty agametic regions between stem cells and cystocyte clusters (as revealed by Orb staining; see Figure 5) within the germarium of 2- to 3-day-old nos mutants argues that stem cells are, indeed, specified in nos mutants; however, they fail to generate cystoblasts in an uninterrupted manner.

What is the relationship between Nanos and plasma membrane maintenance in the germline?
The EM examination of the nos mutant germaria (Figure 7 and Figure 8) indicates that the plasma membrane of germ cells degenerates in nos mutants. This could also explain why stem cells cease to function with age in nos mutants, that is, that they begin to suffer from membrane degeneration and die. Questions arise as to why the plasma membrane of germ cells within the germarium degenerates in nos mutants and what is the relationship between nos and plasma membrane integrity. Biological membranes do not form de novo but grow by expansion of existing membranes. Lipids and proteins are first inserted into preexisting membranous elements in the rough endoplasmic reticulum (ER). The new membrane then flows through the smooth ER and the Golgi to reach the plasma membrane. The membrane recycling (i.e., endocytosis and exocytosis) is thus responsible for a constant turnover of the plasma membrane. One possibility is that the loss of nos activity indirectly affects the plasma membrane recycling. Because Nos is an RNA-binding protein and has been shown to regulate translation of the hunchback or bicoid mRNAs during embryogenesis (TAUTZ 1988 ; STRUHL 1989 ; WANG and LEHMANN 1991 ), it could regulate the expression of a protein required for membrane recycling. Alternatively, a progressive loss of a nos-regulated membrane protein(s) required for maintaining the structural integrity of the plasma membrane could adversely affect the stability of the membrane, leading to its degeneration. Nonetheless, these results indicate a requirement for nos in the maintenance of the plasma membrane of germ cells in Drosophila.

Mitochondrial hyper-proliferation in the germaria of nanos mutants:
In wild type, mitochondrial proliferation for maternal deposition normally occurs in nurse cells beginning at stage 1 of egg chamber development, and the developing germ cells within the germarium contain only a few mitochondria per cell. In ovaries from nos mutants that are 6–7 days old, the germarium was found to be filled with mitochondria (Figure 8, A–C, long arrows). While this mitochondrial phenotype was germ cell dependent (see Figure 8D), the generation of excessive mitochondria appears to have occurred only after germ cell death and within the germarium. This interpretation is consistent with the previous results in sea urchin embryos. When sea urchin embryos were divided into nucleated and enucleated halves, a mitochondrial hyper-proliferation was observed only in the enucleated half, not in the nucleated half (RINALDI et al. 1979 ; see ATTARDI and SCHATZ 1988 ). Thus, these results suggest that certain nuclear-encoded factors repress mitochondrial biogenesis and that removal of the nucleus either through physical means or due to death of germ cells, as in the case of the nos mutants, removes this repressive effect. This inverse relation between nos and mitochondrial biogenesis in the germarium is somewhat intriguing for two reasons. First, the Xcat-2 gene of Xenopus, which has sequence homology with nos (MOSQUERRA et al. 1993 ), was found to be associated with mitochondrial "cloud" in embryos (FORRISTALL et al. 1995 ; KLOC and ETKIN 1995 ), although the functional significance of this association of Xcat2 and mitochondrial cloud has not been determined. Second, in Drosophila nos RNA is found to be associated with polar granules of the pole plasm, which are enriched with mitochondria. The exact relationship between nos and maintenance of membrane integrity or nos and mitochondrial biogenesis activity remains to be determined.

	ACKNOWLEDGMENTS

I would like to express my thanks to Dr. Paul Schedl for his support, enthusiasm, interest, and very valuable suggestions during the course of my work in his lab. Nearly half of the work presented in this article was done in Dr. Schedl's lab at Princeton University. The gift of nanos alleles and the nanos antibody from Dr. Ruth Lehmann is very much appreciated. I also thank her for suggestions and comments on the manuscript. I acknowledge the very valuable help from Felice Farber and Joe Goodhouse in doing the electron microscopy of the specimens and Dr. David Glover for the cyclin antibody. Thanks are due to Drs. Daniel Bopp, Valerie Lantz, Paul Lasko, and Carl Wu for the Sxl, Orb, Vasa, and GAGA antibodies, respectively, and to members of the Bhat lab for comments. This work is supported in part by a grant from the National Institutes of Health to K.B. (R01GM58237), and grants from NIH to Paul Schedl. Financial support from The Human Frontier Science Program Organization and the Department of Cell Biology, Emory University, is gratefully acknowledged.

Manuscript received July 8, 1998; Accepted for publication December 22, 1998.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ATTARDI, G. and G. SCHATZ, 1988  Biogenesis of mitochondria. Annu. Rev. Cell Biol. 4:289-333.

BHAT, K. M. and P. SCHEDL, 1997a  Establishment of stem cell identity in the Drosophila germline. Dev. Dyn. 210:371-382[Medline].

BHAT, K. M. and P. SCHEDL, 1997b  Requirements of engrailed and invected genes reveal novel regulatory interactions between engrailed/invected, patched, gooseberry and wingless during Drosophila neurogenesis. Development 124:1675-1688[Abstract].

BHAT, K. M., G. FARKAS, F. KARCH, H. GYURKOVICHS, and J. GAUSZ et al., 1996  The GAGA factor is required in the early Drosophila embryo not only for transcriptional regulation but also for nuclear division. Development 122:1113-1124[Abstract].

BOPP, D., J. I. HORABIN, R. A. LERSCH, T. W. CLINE, and P. SCHEDL, 1993  Expression of the Sex-lethal gene is controlled at multiple levels during Drosophila oogenesis. Development 118:797-812[Abstract].

BROWN, E. H. and R. C. KING, 1964  Studies on the events resulting in the formation of an egg chamber in Drosophila melanogaster. Growth 28:41-81[Medline].

CURTIS, D., A. APFELD, and R. LEHMANN, 1995  nanos is an evolutionarily conserved organizer of anterior-posterior polarity. Development 121:1899-1910[Abstract].

DALBY, B. and D. M. GLOVER, 1993  Discrete sequence elements control posterior pole accumulation and translational repression of maternal cyclin B RNA in Drosophila. EMBO J. 12:1219-1227[Medline].

DEXTER, T. M. and E. SPOONCER, 1987  Growth and differentiation in the haemopoietic system. Annu. Rev. Cell Biol. 3:423-441.

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].

FORRISTALL, C., M. PONDEL, L. CHEN, and M. L. KING, 1995  Patterns of localization and cytoskeletal association of two vegetally localized RNAs, VG1 and Xcat-2. Development 121:201-208[Abstract].

GAVIS, E. and R. LEHMANN, 1992  Localization of nanos RNA controls embryonic polarity. Cell 71:301-313[Medline].

HALL, P. A. and F. M. WATT, 1989  Stem cells: the generation and maintenance of cellular diversity. Development 106:619-633[Free Full Text].

HULSKAMP, M., C. SCHRODER, C. PFEIFLE, H. JACKEL, and D. TAUTZ, 1989  Posterior segmentation of the Drosophila embryo in the absence of a maternal posterior organizer gene. Nature 338:629-632[Medline].

HULSKAMP, M., C. PFEIFLE, and D. TAUTZ, 1990  A morphogenic gradient of hunchback protein organizes the expression of the gap genes Kruppel and knirps in the early Drosophila embryo. Nature 346:577-580[Medline].

IRISH, V., R. LEHMANN, and M. AKAM, 1989  The Drosophila posterior-group gene nanos functions by repressing hunchback activity. Nature 338:646-648[Medline].

KING, R. C., 1970 Ovarian Development in Drosophila melanogaster. Academic Press, New York.

KLOC, M. and L. ETKIN, 1995  Two distinct pathways for the localization of RNAs at the vegetal cortex in Xenopus oocytes. Development 121:287-297[Abstract].

KOBAYASHI, S., M. YAMADA, M. ASAOKA, and T. KITAMURA, 1996  Essential role of the posterior morphogen nanos for germline development in Drosophila. Nature 380:708-711[Medline].

KOCH, E. A. and R. C. KING, 1966  The origin and early differentiation of the egg chamber of Drosophila melanogaster. J. Morphol. 119:283-304[Medline].

LANTZ, V., L. AMBROSIO, and P. SCHEDL, 1992  The Drosophila orb gene is predicted to encode sex-specific germline RNA-binding proteins and has localized transcripts in ovaries and early embryos. Development 115:75-88[Abstract].

LANTZ, V., J. S. CHANG, J. I. HORABIN, D. BOPP, and P. SCHEDL, 1994  The Drosophila orb RNA-binding protein is required for the formation of the egg chamber and establishment of polarity. Genes Dev. 8:598-613[Abstract/Free Full Text].

LASKO, P. F. and M. ASHBURNER, 1990  Posterior localization of vasa protein correlates with, but is not sufficient for, pole cell development. Genes Dev. 4:905-921[Abstract/Free Full Text].

LEHMANN, R. and C. NUSSLEIN-VOLHARD, 1991  The maternal gene nanos has a central role in pattern formation of the Drosophila embryo. Development 112:679-691[Abstract].

LIN, H. and A. C. SPRADLING, 1993  Germline stem cell division and egg chamber development in transplanted Drosophila germaria. Dev. Biol. 159:140-152[Medline].

LIN, H. and A. C. SPRADLING, 1997  A novel group of pumilio mutations affects the asymmetric division of germline stem cells in the Drosophila ovary. Development 124:2463-2476[Abstract].

MAHOWALD, A. P., and M. P. KAMBYSELLIS, 1980 Oogenesis, pp. 141–224 in The Genetics and Biology of Drosophila, edited by M. ASHBURNER and T. R. F. WRIGHT. Academic Press, New York.

MCKEARIN, D. and A. C. SPRADLING, 1990  bag of marbles: a Drosophila gene required to initiate male and female gametogenesis. Genes Dev. 4:2242-2251[Abstract/Free Full Text].

MOSQUERRA, L., C. FORRISTALL, Y. ZHOU, and M. L. KING, 1993  A mRNA localized to the vegetal cortex of Xenopus oocytes encodes a protein with a nanos-like zinc finger domain. Development 117:377-386[Abstract].

RINALDI, A. M., D. G. LEO, A. ARZONE, I. SALCHER, A. STORACE et al., 1979 Proc. Natl. Acad. Sci. USA 76: 1916–1920.

SCHUPBACH, T. and E. WIESCHAUS, 1989  Female sterile mutations on the second chromosome of Drosophila melanogaster. Genetics 121:101-117[Abstract/Free Full Text].

SCHUPBACH, T. and E. WIESCHAUS, 1991  Female sterile mutations on the second chromosome of Drosophila melanogaster. Genetics 129:1119-1136[Abstract].

STRUHL, G., 1989  Differing strategies for organizing anterior and posterior body pattern in Drosophila embryos. Nature 338:741-744[Medline].

STRUHL, G., P. JOHNSTON, and P. A. LAWRENCE, 1992  Control of Drosophila body pattern by the hunchback morphogen gradient. Cell 69:237-249[Medline].

T