Throughout their lives, adult Drosophila females continuously produce oocytes, each surrounded by an epithelial monolayer of follicle cells. To characterize the somatic stem cells that give rise to ovarian follicle cells, we marked dividing cells using FLP-catalyzed mitotic recombination and analyzed the resulting clones. Each ovariole in young females contains, on average, two somatic stem cells located near the border of germarium regions 2a and 2b. The somatic stem cells do not coordinate their divisions either with each other or with the germline stem cells. As females age, initially mosaic ovarioles become monoclonal, indicating that functional somatic stem cells have a finite life span. Analysis of agametic flies revealed that somatic cells continue to divide in the absence of a germline. Under these conditions, the somatic stem cells develop near the tip of the ovariole (the normal site of the germline stem cells), and a subpopulation of somatic cells that normally separates the germline and somatic stem cells is missing.

Stem cells in adult animals maintain cell populations that are continually shed or turned over. Stem cells have proven difficult to identify and study due to their rarity and lability in culture (see Potten and Loeffler, 1990), but some insights into stem cell regulation have begun to emerge. For example, external signals impinging on specific cellular receptors control the proliferation and differentiation of hematopoietic stem cells (Clark and Kamen, 1987; Metcalf, 1989). A signal from the somatic distal tip cell is required for the continued proliferation of nearby germline cells in the nematode ovary (Kimble et al., 1992). Mechanisms maintaining stem cells probably resemble those used to control unequal or reiterative divisions in developing embryos. Genes required for these differential embryonic cell divisions have been identified (Horvitz, 1988; Umeura et al., 1989; Doe et al., 1991; Herman and Horvitz, 1994); at least one encodes a protein distributed unequally between daughter cells (Rhyu et al., 1994).

The Drosophila ovary provides an especially favorable system for analyzing two distinct groups of stem cells that remain active during much of adult life. Within each ovariole of the ovary, new cysts of 16 interconnected germline cells are enveloped with monolayers of epithelial follicle cells in an anterior region called the germarium (Fig. 1A), and these egg chambers subsequently develop into mature eggs over the course of approximately 5 days (reviewed by King, 1970; Mahowald and Kambysellis, 1980; Spradling, 1993). 2-3 germline stem cells have been identified near the anterior tip of the germarium (Koch and King, 1966, Carpenter, 1975). Their number, location and function as stem cells have been confirmed by clonal analysis (Schupbach et al., 1978; Wieschaus and Szabad, 1979; Wieschaus et al., 1981) and laser ablation (Lin and Spradling, 1993). Germline stem cell daughters undergo four synchronous rounds of division to produce new cysts within the anterior germarial zone known as region 1 (Fig. 1). One cyst cell will become the oocyte, while the others develop into the trophic nurse cells. After traversing region 2a, cysts become surrounded by somatic follicle cells that begin to migrate in from the surface of the germarium in region 2b. Egg chambers subsequently bud off from the germarium, although they remain connected to adjacent chambers by short stalks of specialized follicle cells (see Ruohola et al., 1991; Cummings and Cronmiller, 1994).

Fig. 1.

(A) Schematic diagram of the germarium (adapted from King, 1970). Anterior is to the left. The germarium contains 4 groups of somatic cells based on the work presented here and previous studies: the terminal filament cells (tf), terminal filament base cells (tfb), the inner sheath cells (isc) which surround regions 1 and 2a, and the follicle cells (fc), which surround and separate each germline cyst beginning in region 2b (fcp). These subpopulations are highlighted using different colors. In addition, various germline cell types including the stem cells (gsc), region 1 cysts, region 2a cysts and the single region 3 cyst are depicted. (B) The system used to mark mitotic clones (Harrison and Perrimon, 1993). A constitutive reporter transgene is reconstituted by heat shock induced FLP-mediated recombination between inactive alleles of the transgene. The daughter cell that receives the active copy is dominantly marked by the reporter as are all its progeny.

Fig. 1.

(A) Schematic diagram of the germarium (adapted from King, 1970). Anterior is to the left. The germarium contains 4 groups of somatic cells based on the work presented here and previous studies: the terminal filament cells (tf), terminal filament base cells (tfb), the inner sheath cells (isc) which surround regions 1 and 2a, and the follicle cells (fc), which surround and separate each germline cyst beginning in region 2b (fcp). These subpopulations are highlighted using different colors. In addition, various germline cell types including the stem cells (gsc), region 1 cysts, region 2a cysts and the single region 3 cyst are depicted. (B) The system used to mark mitotic clones (Harrison and Perrimon, 1993). A constitutive reporter transgene is reconstituted by heat shock induced FLP-mediated recombination between inactive alleles of the transgene. The daughter cell that receives the active copy is dominantly marked by the reporter as are all its progeny.

A second group of generative cells must also be present in the germarium to produce a continuous supply of follicle cells. Unlike the germline stem cells, however, somatic stem cells have not been recognized by a distinctive morphology or location. In this respect, ovarian somatic cells closely resemble most studied vertebrate tissues. Somatic cells in region 1 and 2a were rarely observed to label with [3H]thymidine (Chandley, 1966; Carpenter, 1981), suggesting that somatic stem cells reside more posteriorly. Wieschaus et al. (1981) marked X-ray induced follicle cell clones using a mutation affecting eggshell structure. Their results suggested that each ovariole contains multiple somatic stem cells. However, the low frequency of mitotic clones, the partially non-autonomous nature of the marker and the fact that only mature egg chambers could be scored for the presence of clones, prevented further conclusions.

We have marked clones of both germline and somatic ovarian cells using FLP-catalyzed site-specific recombination (Golic and Lindquist, 1989; Harrison and Perrimon, 1993). These studies have identified the number and location of the somatic stem cells, and provided some early insights into their behavior and regulation.

Strains and fly culture

The X-15-29, X-15-33 and hs-FLP stocks used for clonal analysis (Harrison and Perrimon, 1993) were provided by D. Harrison. In these constructs, the lacZ gene is driven by the α84BTubulin promoter, which is expressed at high levels in ovarian cells (Matthews et al., 1989). Other stocks and markers are described in FlyBase (1994). Enhancer trap lines PZ2954, PZ1444 and PZ9383 bearing the PZ transposon were identified in a single P element mutagenesis screen (Karpen and Spradling, 1992; Spradling, 1993). All stocks were maintained at 25°C on standard cornmeal/molasses/agar media.

Developmental staging

Morphological definitions of the stages of oogenesis followed those of King (1970). The best values for the durations of these stages in hours are those determined by Lin and Spradling, (1993).

Generating clones

We generated mitotic clones by the method of Harrison and Perrimon (1993). Females of the genotype X-15-33/X-15-29; MKRS, hs-FLP/+ were produced by standard crosses. To induce FLP expression, culture vials of well-fed, adult flies 1-3 days after eclosion were immersed in a circulating 37°C water bath. Flies were transferred daily to fresh, yeasted food to encourage females to lay continuously. Clonal patches were identified by the expression of the β-galactosidase reporter as detected by either enzyme activity or antibody staining. At the moderate clone induction frequencies used in these experiments, follicle cell clones could be easily identified as discrete clusters of labelled cells. By reducing the length of heat shock, we confirmed that these clusters were independent clones.

β-galactosidase staining

After dissection in EBR (see Ashburner, 1993), the ovaries were fixed in 15 mM KH2PO4-K2HPO4 (pH 6.8), 75 mM KCl, 25 mM NaCl, 3.3 mM MgCl2, 0.5% glutaraldehyde for 8 minutes with constant agitation and then washed extensively in PBT (1× PBS, O.1% Triton X-100). Staining was carried out at 37°C in 10 mM NaH2PO4Na2HPO4 (pH 7.2), 150 mM NaCl, 1 mM MgCl2.6H2O, 3 mM K4[FeII(CN)6], 3 mM K3[FeIII(CN)6], 0.5% Triton-X-100, 0.2% X-gal.

Antibody staining

Ovaries were dissected in EBR, fixed for 15 minutes with constant agitation in 15 mM KH2PO4-K2HPO4 (pH 6.8), 75 mM KCl, 25 mM NaCl, 3.3 mM MgCl2, 6.2% formaldehyde and then washed in PBT. The tissue was pretreated with 0.3% hydrogen peroxide in methanol for 30 minutes before overnight incubation at room temperature in a 1:3000 dilution of rabbit anti-β galactosidase antibody (Cappel #55976) in PBT with 0.2% bovine serum albumin followed by 2 hours in a 1:500 dilution of biotinylated secondary antibody (Vector ABC Elite kit) in PBT and then 1-2 hours in a 1:5 dilution of ABC reagent (Vector ABC Elite kit) prepared according to the manufacturer’s instructions. Finally, horseradish peroxidase activity was detected by nickel-enhanced diaminobenzidine staining (Patel, 1994).

Tissue was routinely mounted and examined as described by Lin and Spradling (1993). To count follicle cell clones, the egg chambers of individual ovarioles from which the sheath had been removed before fixation were mounted in Aqua Poly Mount (Polysciences) and flattened under a weight. The upper and lower surfaces were then photographed with Nomarski optics on color print film and the labeled cells counted on the pictures. The fraction of the ovariole marked was calculated by dividing the number of labeled cells in each postmitotic egg chamber (i.e., stage 7 or later) by the average number of follicle cells per egg chamber (651 cells/chamber, s.d.=37, n=11). The latter value was obtained by counting the number of follicle cells in chambers in which every cell was marked.

BrdU labeling

Ovaries were dissected and teased apart in EBR, then incubated for 1 hour in EBR containing 10 mM BrdU (Sigma). After washing out the unincorporated nucleotide, the tissue was fixed for antibody staining and peroxide treated as above, followed by 30 minutes in 2 N HCl and 2 minutes in 100 mM borax (sodium tetraborate). The tissue was then immunostained as above using a 1:20 dilution of anti-BrdU monoclonal antibody (Becton-Dickinson #7580).

Calculations

To show that FLP-catalyzed recombination in germline stem cells occurs independently of recombination in somatic stem cells, we measured the frequencies of germline and somatically marked ovarioles 9–11 days AHS (35% and 31%, respectively; n=201 ovarioles). Ovarioles containing a labeled stem cell of both types occurred at a frequency of 11.4% in this sample, a value not significantly different from that expected if labeling were independent (0.35×0.31 = 10.8%).

Experimental design

To mark clones of ovarian cells, we used the method of Harrison and Perrimon (1993) to label mitotic cells in vivo (see Methods). This technique uses a heat-shock-inducible FLP recombinase (Golic and Lindquist, 1989) to produce an active α84BTubulin-lacZ reporter gene fusion by driving recombination between complementary inactive alleles, thus dominantly marking recombinant cells and their progeny (Fig. 1B). Clones are clearly marked in both the soma and the germline, allowing these two populations to be compared in a single experiment. In the embryo, only actively cycling cells appear to be susceptible to FLP-mediated recombination (Harrison and Perrimon, 1993), so quiescent cells in the ovary should not contribute to the observed results. Finally, as shown below, there is no detectable background recombination with this particular system; in the absence of heat shock, no clones were observed.

We expected to identify stem cell clones based on the fact that stem cells are the only dividing cells that remain in the germarium indefinitely. Because the vast majority of dividing cells in an ovariole are not stem cells but committed cells that are proliferating prior to their terminal differentiation, most recombination events will occur in these cells. The clones resulting from these events should be transient, however, since the egg chambers that they belong to will move down the ovariole to be laid as eggs. The normal transit time from the division of a germline stem cell to mature egg is about 8.5 days (King, 1970). Therefore, barring delays or temporary arrests in egg chamber development, at times longer than 8.5 days after heat shock (AHS), transient clones should be cleared from the ovary. Only marked stem cells would be expected to remain in the germarium and continue to divide. Thus, marked cells present in the ovary significantly longer than 8.5 days AHS should be stem cells and their progeny.

Properties of transient follicle cell clones

To confirm that the clone marking system had the expected properties, we sampled stage 10 egg chambers at regular intervals after a 60 minute heat shock and measured the size and frequency of transient follicle cell clones (Figs 2, 3). Follicle cells stop dividing at the end of stage 6, 24 hours before stage 10. By examining a fixed, easily scored stage whose follicle cells have ceased division, we were able to compare egg chambers in which clones had been induced at progressively earlier stages of oogenesis (Table 1). At 1.5 days AHS, chambers had an average of 16 clones with a mean size of 1.6 cells. Over the next 4 days, clone frequency fell to 0.3 clones/chamber while the average clone size rose to 142 cells (Fig. 2A). The inverse relationship between clone size and frequency was expected. Young egg chambers contain fewer follicle cells and are therefore less likely to host a recombination event. The rare cells that are marked in young chambers have more divisions remaining before stage 10, resulting in larger clones.

Table 1.

Quantitative analyisis of transient clones

Quantitative analyisis of transient clones
Quantitative analyisis of transient clones
Fig. 2.

Kinetics of clonal labeling. (A) Changes in clone size and clone frequency in stage 10 egg chambers with increasing time after a 60 minute heat shock. From the slope of the best fit line to time versus log(clone size), we calculated a doubling time for follicle cells of 9.6 hours.(B) Decreasing number of precursor follicle cells per chamber as a function of increasing time after heat shock (see Table 1). The number of follicle cell precursors/chamber plateaus between 2 and 3, suggesting that each somatic stem cell contributes once to each egg chamber.

Fig. 2.

Kinetics of clonal labeling. (A) Changes in clone size and clone frequency in stage 10 egg chambers with increasing time after a 60 minute heat shock. From the slope of the best fit line to time versus log(clone size), we calculated a doubling time for follicle cells of 9.6 hours.(B) Decreasing number of precursor follicle cells per chamber as a function of increasing time after heat shock (see Table 1). The number of follicle cell precursors/chamber plateaus between 2 and 3, suggesting that each somatic stem cell contributes once to each egg chamber.

Fig. 3.

Size and location of transient clones. Ovaries were stained for βgalactosidase activity a variable number of days after heat shock (AHS) to induce FRT-mediated recombination. (A-G) Transient clones in stage 10 egg chambers at(A) 1.5 days AHS, (B) 2 days AHS, (C) 2.5 days AHS, (D) 3 days AHS, (E) 4 days AHS. (F) A putative polar cell clone spanning both ends of an egg chamber, 3 days AHS.(G) Another small clone which appears to label both posterior polar cells (inset shows labeled pair of cells at higher magnification).(H) Stage 12 egg chamber 4 days AHS showing mosaicism among the follicle cells forming the dorsal appendages. The arrowhead indicates a group of unlabelled follicle cells. The scale bar in A applies to A-G.

Fig. 3.

Size and location of transient clones. Ovaries were stained for βgalactosidase activity a variable number of days after heat shock (AHS) to induce FRT-mediated recombination. (A-G) Transient clones in stage 10 egg chambers at(A) 1.5 days AHS, (B) 2 days AHS, (C) 2.5 days AHS, (D) 3 days AHS, (E) 4 days AHS. (F) A putative polar cell clone spanning both ends of an egg chamber, 3 days AHS.(G) Another small clone which appears to label both posterior polar cells (inset shows labeled pair of cells at higher magnification).(H) Stage 12 egg chamber 4 days AHS showing mosaicism among the follicle cells forming the dorsal appendages. The arrowhead indicates a group of unlabelled follicle cells. The scale bar in A applies to A-G.

These experiments validated several assumptions underlying this clone marking system. First, clone induction is limited to a short period following heat shock. From the heterogeneity that they observed in clone size, Harrison and Perrimon (1993) concluded that the FLP recombinase continues to catalyze recombination for several hours after induction. On the time scale of our experiments, however, perdurance of the recombinase was negligible since the clone size increased continuously in the first few days after heat shock and small clones did not continue to arise at later times. Second, the heat shock used to induce FLP protein production does not cause any significant developmental arrest. Even the most severe heat shock used in our studies (60 minutes at 37°C) did not alter the distribution of egg chamber stages within ovarioles (an indication of developmental arrest) or detectably retard the expected rate at which marked chambers progressed to increasingly mature developmental stages. Third, the labeling frequency (Table 1) did not change more than 2–3 fold throughout oogenesis (5.8%19%). Finally, as expected, transient germline clones were also observed (Fig. 3C,E,G).

Several useful facts about follicle cell development could also be deduced from these experiments. First, follicle cells of different lineages do not mix on the surface of the egg chamber. Although the borders of a clone can be quite irregular, the cells of a clone formed a coherent unit (Fig. 3). Second, our data on clone size allowed us to calculate a doubling time for follicle cells of 9.6 hours over at least 3 days (Fig. 2A). Third, we were able to calculate the number of follicle cells per egg chamber as a function of time before stage 10 (see Table 1). The numbers deduced from these calculations (Table 1) agree closely with the known biology of follicle cell development. For example, we calculate that an egg chamber reaching stage 10 at 2 days AHS contained 78 precursor cells at the time of heat shock, when the cyst was entering region 3 of the germarium. Budding cysts in region 3 contain about 80 follicle cells (Mahowald and Kambysellis, 1980).

Lineage restrictions in the follicle cell layer

These clones also allow us to determine what role cell lineage plays in the development of the follicle cell layer. First, there was no correlation to egg chambers. We observed clones that extended between adjacent chambers and included the interconnecting stalk cells. Within egg chambers, subsets of follicle cells undergo specific migrations and synthesize particular chorionic structures, such as the dorsal appendages (see Spradling, 1993). Any such population that is clonally derived from one of the founder follicle cells would never be split by a clone border. We found no evidence for lineage restrictions among the characterized subpopulations of follicle cells. For example, the clone in Fig. 3G, spans follicle cells that remain over the nurse cells and those that migrate centripetally to form the anterior chorion. The clone in Fig. 3H includes both main body follicle cells and a subset of those forming dorsal appendages.

A special pair of ‘polar’ follicle cells has been identified at the anterior and posterior pole of each egg chamber by their specific expression of fasciclin III (Brower et al., 1981; Patel et al., 1987), neuralized (Ruohola et al., 1991), and other genes (Spradling, 1993). We found that 10% (15/145) of clones 3 days AHS were 1to 2-cell clones confined to the presumptive polar cells. The majority of these clones (10/15) marked single cells at the anterior or posterior pole, the site of the polar cell pair. The other clones marked a pair of adjacent cells at the same end (2/15, Fig. 3H) or at opposite ends (3/15, Fig. 3F) of the egg chamber. Since the average clone size at these times was 24 cells (Fig. 2A), the marked follicle cells stopped dividing long before those around them. Clones restricted to these presumptive polar cells were not observed at 2 days AHS and only rarely (1/74 clones) at 4 days AHS.

Persistent clones are stem cell clones

Having established that we could reliably label dividing cells with this technique and that oogenesis proceeded normally under these conditions, we began to examine ovaries more than 8.5 days AHS for evidence of stem cell clones. By this time all transient clones should have left the ovary as mature eggs and any persistently marked ovarioles would be candidate stem cell clones. Based on previous studies, we could predict the expected behavior of labelled germline stem cells in this assay. First, in contrast to the transient germline clones seen shortly after heat shock in which individual chambers were mosaic (Fig. 3C), the individual chambers produced by a labelled germline stem cell should stain homogeneously due to their clonal origin following a single stem cell division. Second, because each ovariole contains 2–3 germline stem cells, we would expect an average of one third to one half of egg chambers to be labelled. Finally, because the labeling in a single ovariole should be the product of one marked stem cell, after a milder heat shock, we expected to find fewer marked ovarioles, but these should show the same pattern of labeled cells as before.

When we examined the ovaries of flies 9–11 days AHS, these expectations regarding germline stem cell clones were fully confirmed. Germline stem cells were labeled with high frequency, since 30% of ovarioles contained germline-labeled egg chambers with the expected properties, as detailed below. In these same ovaries we also noted that a high fraction of the ovarioles (also about 30%) contained labeled follicle cells (Fig. 4B), suggesting that they derived from labeled somatic stem cells. These clones typically contained hundreds of cells and extended between adjacent egg chambers. We analyzed these persistent somatic clones and compared them with the germline stem cell clones, to further support their origin from labeled somatic stem cells.

Fig. 4.

Persistent clones result from individual stem cells. (A) The percentage of ovarioles containing persistent clones in the germline (solid dots) and soma (open diamonds) was plotted against length of heat shock. Each data point represents 200 –300 ovarioles pooled from multiple independent experiments. In the absence of heat shock, no marked cells were observed (299 ovarioles scored).(B-D) Comparison of ovarioles with persistent somatic clones 10 days after a heat shock of (B) 60 minutes, (C) 30 minutes, or (D) 15 minutes. No significant differences in the size or appearance of the persistent clones was observed.

Fig. 4.

Persistent clones result from individual stem cells. (A) The percentage of ovarioles containing persistent clones in the germline (solid dots) and soma (open diamonds) was plotted against length of heat shock. Each data point represents 200 –300 ovarioles pooled from multiple independent experiments. In the absence of heat shock, no marked cells were observed (299 ovarioles scored).(B-D) Comparison of ovarioles with persistent somatic clones 10 days after a heat shock of (B) 60 minutes, (C) 30 minutes, or (D) 15 minutes. No significant differences in the size or appearance of the persistent clones was observed.

Stem cell clones should still be present in the ovary long after induction, whereas all other clones should be lost with time, even in ovarioles where egg chamber development was temporarily arrested. We found persistent germline and somatic clones up to 26 days AHS (the longest time point examined). No differences in the size or other properties of either somatic or germline clones were observed in comparing young and old flies. It is extremely unlikely marked cells could persist this long after heat shock unless a generative cell(s) responsible for continued egg chamber production had been labeled.

The persistent somatic cell clones labeled a significant fraction of all the follicle cells in a marked ovariole. This could have resulted from marking one member of a small group of stem cells, or by marking multiple cells within a larger population of progenitor cells. The possibility that multiple cells were marked was not unreasonable, since FLP-mediated recombination must have occurred in a large fraction of the germline stem cells to generate persistent clones in 30% of all ovarioles. However, these two possibilities could be distinguished experimentally. If multiple recombination events in a population of generative somatic cells were responsible for the persistent follicle cell clones, then the number of labeled cells per ovariole should decrease when the overall rate of recombination was reduced by using milder heat shocks. In contrast, if a single stem cell gave rise to the persistently labeled cells, then reducing the overall clone frequency should cause relatively little change in the number of labeled cells per ovariole. When milder heat shocks were carried out, the frequency of ovarioles with persistent somatic or germline clones declined (Fig. 4A), but the general pattern of mosaicism in the remaining marked ovarioles was unchanged (Fig. 4B-D, data for germline not shown). When the heat shock was reduced from 60 to 30 minutes, the fraction of follicle cells labelled in somatically marked ovarioles also remained constant [53% (s.d.=28%, n=16 ovarioles) and 49% (s.d.=24% n=18 ovarioles), respectively] even though the clone frequency dropped by more than two fold. Similarly, the fraction of germline-marked chambers per mosaic ovariole was unchanged when the heat shock was reduced from 60 to 30 minutes [40% (s.d.=15%, n=31 ovarioles) and 39% (s.d.=14%, n=20 ovarioles), respectively]. These data all support the conclusion that persistent clones in both the soma and germline are the product of individual marked stem cells.

Stem cell number

The size and properties of stem cell clones allowed us to determine the number and behavior of these cells. In any clonal marking experiment, the number of progenitor cells present at the time of induction is the reciprocal of the fraction of the tissue constituting the clone. Since on average 50% of the follicle cells in somatically mosaic ovarioles were labeled, each ovariole contains 2 somatic stem cells. In ovarioles with mosaic germlines, 40% of the egg chambers were labeled, so an ovariole contains 2.5 germline stem cells. This latter value agrees well with the previous measurement of 2.8 germline stem cell per ovariole determined by Wieschaus and Szabad (1979).

Coordination of stem cell divisions within an ovariole

The pattern of mosaicism in a marked ovariole provides a history of the recent divisions of the stem cells. For the germline cells, each egg chamber represents one division of a stem cell and germline cysts generally maintain their birth order as they progress through the germarium (Carpenter, 1981). The presence of several consecutively marked egg chambers within an ovariole flanked by unmarked chambers then records a sequence of consecutive divisions (a ‘burst’) by the marked stem cell. We measured 152 bursts of 1–5 egg chambers in 109 ovarioles. 65% contained single egg chambers and 28% contained 2 egg chambers, so individual stem cells did not tend to fire repeatedly. In contrast, Schüpbach et al. (1978) reported that germline stem cells divide preferentially in bursts of 2–4 cells and stop after a single division only 8% of the time. It is not clear why our results differed from those reported previously. Unlike the germline, the individual clonal patches of cells derived from a series of consecutive somatic stem cell divisions lie adjacent to each other and cannot be distinguished. Consequently, we could not determine the burst size of these cells.

In the testis, the division of germline and cyst stem cells is thought to be tightly co-regulated. Following the division of a germline stem cell, an adjacent somatic stem cell is activated and its progeny surround the new germline cyst (Lindsley and Tokuyasu, 1980). To look for a similar coordination in the ovary, we examined ovarioles with persistent clones in both the germline and soma (Fig. 5C and Methods). Ovarioles containing persistent clones in both the germline and the soma showed no correlation between the location of labeled follicle cell patches and marked germline cysts (Fig. 3C). This observation suggests that the divisions of these two classes of stem cells are not coordinately regulated.

Fig. 5.

Independent regulation of somatic and germline stem cells. (A) Persistent somatic clone 10 days AHS. (B) Persistent germline clone, 11 days AHS. (C) Ovariole with clones in both germline and soma showing the absence of coordination between these two stem cell types, 10 days AHS.

Fig. 5.

Independent regulation of somatic and germline stem cells. (A) Persistent somatic clone 10 days AHS. (B) Persistent germline clone, 11 days AHS. (C) Ovariole with clones in both germline and soma showing the absence of coordination between these two stem cell types, 10 days AHS.

Stem cell senescence

The fecundity of Drosophila females declines with age (David et al., 1974). To look for age-associated changes in stem cell function, we compared persistent clones in flies 9–11 days AHS with those in flies 20–22 days AHS. For an ovariole with two stem cells, one marked and the other unmarked, the loss of the marked stem cell would result in an unmarked ovariole, while the loss of the unmarked stem cell would result in a completely marked ovariole. Thus, loss of stem cells should produce a decrease in the number of partially marked ovarioles and an increase in the fraction completely labeled. This is what we observed.

We examined 402 ovarioles 9–11 days AHS and 311 ovarioles at 20–22 days AHS. The frequency of completely marked ovarioles rose from 0.5% to 5.8% for the soma and from 1.7% to 7.2% for the germline (Fig. 6A). At the same time, the frequency of partial or mosaic ovarioles declined from 28% to 11.2% for the soma and from 30% to 8.5% for the germline (Fig. 6A). This result is most simply explained by occasional loss of a stem cell, due to inactivation, death or differentiation. To determine the rate of stem cell loss, we plotted the fraction of mosaic ovarioles (i.e., ovarioles with multiple stem cells) in the population at several times AHS (Fig. 6B). Since flies were 1–3 days old at the time of heat shock, our results showed that about half the stem cells of both types cease functioning between 13 and 29 days after eclosion (Fig. 6).

Fig. 6.

Senescence of stem cell activity. (A) Comparison of the frequency of partial and complete persistent clones at short and long times (9–11 days versus 20–22 days) after a 60 minute heat shock. These data are the average of 4 separate experiments. (B) The percentage of ovarioles with partial clones at increasing times AHS for one of the experiments included in A showing the loss of stem cells with time.

Fig. 6.

Senescence of stem cell activity. (A) Comparison of the frequency of partial and complete persistent clones at short and long times (9–11 days versus 20–22 days) after a 60 minute heat shock. These data are the average of 4 separate experiments. (B) The percentage of ovarioles with partial clones at increasing times AHS for one of the experiments included in A showing the loss of stem cells with time.

Locating the somatic stem cells

Despite their activity, it has not previously been possible to recognize the somatic stem cells or localize them within the ovariole. However, the clonal marking system used in our experiments provides a direct approach to this problem. For each clone, the most apical marked cell should represent the stem cell. This expectation was confirmed in the case of germline clones, as could be most easily seen in the case of fully marked ovarioles. In ovarioles with complete germline clones, labeling always extended to the tip of the germarium and could be clearly seen to include a stem cell (Fig. 7A,B). In the germaria of ovarioles with complete somatic clones, however, labeling extended only to the border of regions 2a and 2b, suggesting that the stem cells lie in this zone (Fig. 7C,D). Ovarioles with partial somatic clones 9 –10 days AHS exhibited the same anterior limit of clonal labeling.

Fig. 7.

Localization of stem cells. Ovarioles with complete clones were immunostained for βgalactosidase and the position of the anterior-most labeled cell determined. (A,B) Complete germline clone at two magnifications showing labeling of every germline cell, including a single cell just below the terminal filament, the known location of germline stem cells. (C,D) Complete somatic clone at two magnifications showing labeling of all somatic cells posterior of the region 2a/2b border. The anterior margin of the germarium in D is indicated by a dotted line.

Fig. 7.

Localization of stem cells. Ovarioles with complete clones were immunostained for βgalactosidase and the position of the anterior-most labeled cell determined. (A,B) Complete germline clone at two magnifications showing labeling of every germline cell, including a single cell just below the terminal filament, the known location of germline stem cells. (C,D) Complete somatic clone at two magnifications showing labeling of all somatic cells posterior of the region 2a/2b border. The anterior margin of the germarium in D is indicated by a dotted line.

Frequently, a few somatic cells were labeled more anteriorly in the germarium, in regions 1-2a. However, these cells appeared to be unrelated to follicle cell progenitors. First, only individual, sporadically distributed cells were labeled, rather than contiguous clones. Second, they were present at equal frequency in ovarioles that lacked labeling outside the germarium. Third, their frequency declined with decreasing heat shock, while the properties of persistent somatic clones was unchanged. Therefore, these labeled cells appeared to result from recombination in non-cycling cells that were unrelated to follicle cell progenitors.

To confirm that the somatic stem cells are located part way down the germarium near the region 2a/2b border, we mapped replicative cells in the germarium by BrdU incorporation (see Methods). If somatic stem cells were actually located anterior to region 2b but were unable to express the αTubulin-lacZ gene, we would expect to find cycling cells overlying region 1 or 2a. If the sporadically marked region 1-2a cells were part of a stream of developing prefollicle cells, then either they or more anterior cells should be actively cycling. Contrary to the predictions of these models, no somatic cells anterior to region 2b were labeled following a 1 hour incubation with BrdU, despite heavy and consistent labeling for more posterior somatic cells (Fig. 8A). These results strongly supported the location of the somatic stem cells indicated by clonal analysis.

Fig. 8.

Behavior of germarial somatic cells in the absence of a germline. (A,B) Pattern of BrdU incorporation in wild-type (A) and agametic (B) germaria. The asterisks in A indicate the border of regions 2a and 2b. Labeling anterior to this point is restricted to germline cysts. (C-H) Comparison of enhancer trap expression patterns in normal (C,E,G) and agametic (D,F,H) germaria. (C,D) The expression pattern of the enhancer trap line PZ9383 shows that the terminal filament is unaffected in the agametic mutant. (E,F) Enhancer trap line PZ1444 is normally expressed in the cells at the base of the terminal filament and the somatic cells of regions 1 and 2a (E). In the agametic ovary, only the expression at the base of the terminal filament remains (F). (G,H) Enhancer trap line PZ2954 is normally expressed in all germarial follicles from region 2b onwards; however, in the agametic germarium, expression commences just posterior of the base of the terminal filament. The arrowheads in A-H indicate the base of the terminal filament as a conserved landmark in each of these germaria.

Fig. 8.

Behavior of germarial somatic cells in the absence of a germline. (A,B) Pattern of BrdU incorporation in wild-type (A) and agametic (B) germaria. The asterisks in A indicate the border of regions 2a and 2b. Labeling anterior to this point is restricted to germline cysts. (C-H) Comparison of enhancer trap expression patterns in normal (C,E,G) and agametic (D,F,H) germaria. (C,D) The expression pattern of the enhancer trap line PZ9383 shows that the terminal filament is unaffected in the agametic mutant. (E,F) Enhancer trap line PZ1444 is normally expressed in the cells at the base of the terminal filament and the somatic cells of regions 1 and 2a (E). In the agametic ovary, only the expression at the base of the terminal filament remains (F). (G,H) Enhancer trap line PZ2954 is normally expressed in all germarial follicles from region 2b onwards; however, in the agametic germarium, expression commences just posterior of the base of the terminal filament. The arrowheads in A-H indicate the base of the terminal filament as a conserved landmark in each of these germaria.

Effect of germline on somatic stem cells

Germline and somatic cells communicate during egg chamber development (Goode et al., 1992). To study whether the germline influences the activity of somatic stem cells, we analyzed somatic cell proliferation in flies that lacked germline cells. Agametic flies have very small ovaries in which each ovariole contains a small mass of disorganized cells surrounded by a thickened sheath (Fielding, 1967). We generated agametic flies for our experiments using the maternal effect, temperature-sensitive mutation oskar301 (Lehmann and Nüsslein-Volhard, 1986). Except for their lack of germ cells, the daughters of osk301 homozygous females kept at 18°C develop normally. We first identified growing cells in agametic ovaries by labeling S phase cells using BrdU (Fig. 8B). 2–20 cells were labeled by BrdU; however, unlike the wild type, these cells lay almost adjacent to the terminal filament. The mass of somatic cells within these ovarioles, whose size increases in older females, is apparently produced by dividing cells confined to the anterior end of the germarium.

To understand the altered location of proliferative somatic cells in agametic compared to wild-type ovarioles, we needed markers to identify the other cells present. Enhancer trap lines were identified that express β-galactosidase in subsets of somatic cells (see Methods). Four subsets of somatic cells in the wild-type germarium were recognized based on these expression patterns: the terminal filament, cells at the base of the terminal filament, cells surrounding region 1 and 2a, and cells expressed in 2b and 3 (Figs 1, 8). Each marker was crossed to osk301 females and its pattern of expression determined in the ovaries of female progeny. For each subset of cells, 3–4 independent lines expressed in these cells were tested in the agametic flies.

The results of these experiments suggested that the somatic stem cells develop more anteriorly in the absence of developing germline cells. The anterior ends of the agametic ovarioles appeared normal, and marker expression in the terminal filament was unchanged (Fig. 8C,D). In wild-type germaria, 20–40 cells located over region 1 –2a cells are specifically labeled by certain enhancer trap lines; however, these cells appeared to be absent in agametic germaria. For example, the line shown (Fig. 8E) labels cells at the base of the terminal filament and in region 1 –2a; in agametic germaria only the terminal filament base cells continued to express β-galactosidase (Fig. 8F). Consistent with the absence of cells with a region 1–2a identity, region 2b-3-specific markers (Fig. 8G) were expressed in all cells posterior to the base cells (Fig. 8H). Multiple independent enhancer trap lines exhibited the same shift in cell position in agametic ovaries. Thus, the altered location of the proliferating somatic cells near the terminal filament was associated with the absence of a population of somatic cells that are present in the wild-type germarium. This germline-dependent population may correspond to the ‘inner sheath cells’ that have been recognized in a similar position in several other insect groups (Büning, 1994).

These studies revealed for the first time how many somatic stem cells are present in each germarium and where they are located. Both the low number of stem cells (two) and their position at the region 2a/2b border provided new insights into how egg chamber production is regulated. Germline and somatic cell proliferation in the testis appears to be coordinated by interactions between both populations of stem cells and a subpopulation of somatic hub cells, all of which reside close together at the apical tip (Lindsley and Tokayasu, 1980; Gönczy et al., 1992). Like the testis, ovarioles contain a small population of somatic cells near the germline stem cells and adjacent to potential regulatory cells, the terminal filament (Fig. 1A). Our studies, however, clearly indicate that the somatic stem cells actually lie 20–30 μm away (3–5 cell diameters) at the region 2a/2b border. Furthermore, somatic and germline stem cell divisions are not coordinated, and ovarian stem cell products do not associate to form egg chambers based on lineage. Thus, stem cell regulation is likely to differ significantly in male and female gonads.

The localization of the somatic stem cells is supported by four types of evidence. First, stem cell clones terminated at the region 2a/2b border, as clearly shown in fully labeled ovarioles. Second, few if any somatic cells anterior to this region traverse S phase, based on BrdU incorporation. Third, the kinetics of follicle cell growth showed that follicle cell precursors begin proliferating about 4 days before stage 10 is reached (Fig. 2B). This corresponds to the time required for a chamber to develop from the region 2a/2b border to stage 10, but is significantly less than the 6–7 days needed for a cyst to progress from the tip of the germarium to stage 10. Finally, proliferating somatic cells remain in agametic ovarioles, despite the apparent absence of most region 1–2a follicle cells.

The presence of only two stem cells raises several issues. First, these cells reside at the junction between the inner sheath cells and the zone of mitotically active, inwardly migrating cells. The stem cells must be attached in some way at the junction to prevent their own posterior movement and differentiation into follicle cells. However, no morphologically distinguishable cells with the requisite properties have been reported. Most likely, the somatic stem cells appear very much like their immediate progeny. Second, specific foci of proliferating follicle cells have not been identified morphologically. Perhaps the immediate progeny of the somatic stem cells migrate posteriorly and spread laterally along the wall of the germarium prior to invagination, thereby obscuring their origin from two discrete sites.

Mechanism of stem cell action

Our results rule out the possibility that ovarian somatic cells are maintained by a large population of identical cells that stochastically differentiate and are replaced by the division of a neighboring cell. The predicted patterns of labelled cells from such a ‘germinal epithelium’ would be very similar to what we observed if levels of clone induction approached 50%. However, as the strength of the heat shock was decreased and the total number of cells undergoing recombination declined, this model predicts that the number of labeled cells in a mosaic ovariole would also decline. Such changes were never observed. Consequently, we can be confident that only about two cells per ovariole are used to maintain the follicle cell population.

We cannot distinguish at this point whether somatic stem cells divide differentially or whether their daughter cells simply respond to spatial differences in their microenvironment (see Potten and Loeffler, 1990). This question could best be resolved by identifying cellular components that are distributed asymmetrically during stem cell division, or that impinge differentially on the division products. The germline cystoblast and cystocyte divisions have recently been shown to be physically unequal, due to the asymmetric segregation of the membrane skeleton proteins αspectrin and hu-li tai shao (a Drosophila adducin homolog) present in the fusome (Lin et al., 1994; Lin and Spradling, 1995). No somatic cells containing an unusual organization of these components were observed in the germarium (J. Margolis and A. Spradling, unpublished).

Stem cell senescence

The progressive increase in completely labeled ovarioles that we observed in older females indicates that both germline and somatic stem cells have a finite life span in the adult. The approximate number of stem cell divisions prior to senescence can be calculated from our experiments. Each ovariole produces a maximum of 2 egg chambers per day under ideal nutritional conditions, and there is approximately one somatic stem cell division for each egg chamber produced. Consequently, by the time half of the stem cells have been lost 22 days after egg laying begins, they will have completed up to 44 divisions. Further experiments will be required to learn whether stem cells are lost due to differentiation, inactivation or death.

In Drosophila, female fecundity drops off sharply with increasing age (David et al., 1974). Our experiments suggest that this loss of fecundity is caused, at least in part, by a loss of germline and somatic stem cells. Consistent with this view, many ovarioles in the ovaries of aged females lack developing egg chambers (J. Margolis and A. Spradling, unpublished). These observations support the view that the reduced fraction of mosaic, two-stem cell ovarioles that we measured with increasing age is due to an absolute decrease in the number of functional stem cells, and that this decline continue in ovarioles with only one remaining stem cell.

Stem cells in a variety of systems are lost as a function of increasing age. Van Zant and co-workers (1990) have shown that mouse hematopoetic stem cell senescence is genetically controlled and autonomous to the stem cells themselves.

Evidence suggests that a variety of rat stem cell populations decline with age (Brill et al., 1993; Roholl et al., 1994). Explanted human diploid fibroblasts, a model system for studying stem cells, have a restricted life span under typical culture conditions (Goldstein, 1990) due to processes that appear to mimic changes occuring in vivo. The stem cells of mammalian intestinal crypts are also replaced over time, although it is not clear whether the actual number of crypts decreases (Potten and Loeffler, 1990). The Drosophila ovary provides a experimentally tractable system for investigating the molecular basis of stem cell aging. The relatively simple anatomy of the germarium, the short life span, and the availability of sophisticated genetic and molecular tools represent significant advantages over alternative systems.

Differentiation of specialized follicle cell subpopulations

Our experiments show that lineage is not important in defining the subsets of follicle cells that play specialized roles in the development of the egg chamber. We found that groups of follicle cells that undergo specific migrations (including the border cells), remain over the nurse cells, or synthesize specialized portions of the eggshell, generally derive from multiple precursor cells. However, we did find that a small subset of follicle cells near the anterior and posterior poles of the egg chamber cease division long before their neighbors. Because these clones were always located at the extreme anterior and posterior end, and consisted of exactly one or two cells, we presume that they correspond to the ‘polar’ follicle cells that have been described previously.

The specification of polar follicle cells requires Notch and Delta (Ruohola et al., 1991). Based on these results, these authors postulated that polar follicle cells are specified beginning at the time of egg chamber budding in a pathway that also includes stalk cells. Our results suggest that polar cells are actually determined and cease division as cysts move through region 2b, prior to egg chamber budding. Presumptive polar cell clones were recovered only at 3 days following AHS, which corresponds to a time 12–24 hours prior to the onset of budding. Furthermore, clone size data places the number of follicle cell precursors at the time of polar cell clone induction at 14, close to the number of cells that envelope cysts early in region 2b (Carpenter, 1975). Notch and Delta may act to distinguish polar from non-polar cells at this stage, rather than following budding as previously suggested. The early cessation of polar cell growth is consistent with the previous observations that the number of fasicilin III-positive cells does not increase during stages 1–6 when the number of follicle cells in a chamber continues to increase.

Polar cells cease division prior to main body follicle and stalk cell precursors, however, we found no evidence for a strict lineage relationship between these cell groups. Clones induced in other cells at this stage gave rise to patches of about 25 cells that could cover any portion of the egg chamber and the interfollicular stalks. We never observed clones that were restricted to stalks alone, or to stalks and polar cells. Polar cell pairs may develop from a precursor with restricted potential or be determined at this stage as a result of cellular interactions. We could not determine whether chambers with two labeled polar cells resulted from the induction of clones in a specific precursor cell, or from the induction of two independent clones.

Inner sheath cells

Our studies strongly suggested that the cells overlying the developing cysts in region 1–2a constitute a previously unrecognized subset of somatic cells. These cells behaved as a distinct subset based on (1) co-expression of markers in certain enhancer trap lines; (2) failure to incorporate BrdU; (3) absence in agametic ovaries. Ovaries from several insect orders contain a population of quiescent cells at an equivalent position that have been called ‘inner sheath cells’ (Büning, 1994). The structure of the ovariole that emerges from these experiments, in which inner sheath cells separate germline and somatic stem cells, may be shared by a variety of other insect groups.

Studies of germline sex determination have demonstrated that signals from the soma are important for the establishment of germline sexual identity (see Pauli and Mahowald, 1990), however, the timing and location of these signals are yet to be determined. We note that the inner sheath cells constitute a distinct subpopulation that makes contact with germline cells at a time when these signals may be needed. Furthermore, our experiments showed that inner sheath cells do not differentiate in normal numbers in the absence of a germline. It would be worthwhile to directly investigate the role these cells play in germline development.

Regulation of somatic stem cells

How then might the proliferation of somatic stem cells be controlled? About 16 cells initially move in and cover each cyst shortly after entering region 2b (Carpenter, 1981; J. Margolis and A. Spradling, unpublished). To produce these cells, our data suggests that both somatic stem cells divide once, on average, followed by 3 rounds of division of their progeny. We measured a relatively constant rate of 9.6 hr for these divisions (Fig. 2), consistent with the fact that new cysts bud from the germarium approximately every 10–12 hours. Since no actual synchrony is observed between the progeny of marked germline and somatic stem cells (Fig. 5), only a general coordination would be required in order to keep germline and somatic cell populations in relative balance.

Since the germline and somatic stem cells lie several cell diameters apart in the germarium, two general mechanisms might equalize the number of 16-cell cysts and follicle cell precursors that they produce. First, they may directly communicate, for example, by utilizing diffusible signaling molecules. TGF-ß family members can regulate the proliferation of stem cells in a variety of organisms (Podolsky, 1993; Kordon et al., 1995), and modulate cell growth in imaginal disks (Heberlein et al., 1993). Alternatively, coordination between germline and soma might be effected directly. Cysts that contact a somatic stem cell near the region 2a/2b border might induce it to divide by expressing mitogenic signals on their surfaces. Perhaps these same signals cause the main body follicle cells of early cysts to continue dividing, while stalk cells (which do not directly contact the cyst) become quiescent. Membranous contacts between somatic cells and migrating cysts are a striking feature of the normal germarium (King, 1970). This model cannot explain why somatic cells continue to proliferate in agametic ovarioles, however. Whatever the actual mechanism, the Drosophila ovary will provide an excellent system for elucidating how the proliferation of stem cells is molecularly controlled.

We are grateful to Doug Harrison for generating and providing the fly stocks used in this study. We also thank Maggie de Cuevas, Lynne Schneider, and Linda Keyes for their critical reading of this manuscript. J. S. M. was supported by a postdoctoral fellowship from the National Institutes of Health.

Ashburner
,
M.
(
1993
).
Drosophila v.2 A Laboratory Manual
.
Cold Spring Harbor, New York
:
Cold Spring Harbor Press
.
Brill
,
S.
,
Holst
,
P.
,
Sigal
,
S.
,
Zvibel
,
I.
,
Fiorino
,
A.
,
Ochs
,
A.
,
Somasundaran
,
U.
and
Reid
,
L. M.
(
1993
).
Hepatic progenitor populations in embryonic, neonatal and adult liver
.
Proc. Soc. Exptl. Biol. Med
.
204
,
261
269
.
Brower
,
D. L.
,
Smith
,
R. J.
and
Wilcox
,
M.
(
1981
).
Differentiation within the gonads of Drosophila revealed by immunofluorescence
.
J. Embryol. Exp. Morph
.
63
,
233
242
.
Büning
,
J.
(
1994
).
The Insect Ovary
.
London
:
Chapman and Hall
.
Carpenter
,
A. T. C.
(
1975
).
Electron microscopy of meiosis in Drosophila melanogaster females. I. Structure, arrangement and temporal change of the synaptonemal complex in wild-type
.
Chromosoma
51
,
157
182
.
Carpenter
,
A. T. C.
(
1981
).
EM autoradiographic evidence that DNA synthesis occurs at recombination nodules during meiosis in Drosophila melanogaster females
.
Chromosoma
83
,
59
80
.
Chandley
,
A. C.
(
1966
).
Studies on oogenesis in Drosophila melanogaster with 3H-thymidine label
.
Exp. Cell Res
.
44
,
201
215
.
Clark
,
S. C.
and
Kamen
,
R.
(
1987
).
The human hematopoietic colonystimulating factors
.
Science
236
,
1229
1237
.
Cummings
,
C. A.
and
Cronmiller
,
C.
(
1994
).
The daughterless gene functions together with Notch and Delta in the control of ovarian follicle development in Drosophila
.
Development
120
,
381
394
.
David
,
J.
,
Biémont
,
C.
and
Fouillet
,
P.
(
1974
).
Sur la forme des courbes de ponte de Drosophila melanogaster et leur adjustment á des modèles mathématiques
.
Arch. Zool. Exp. Gen
.
115
,
263
277
.
Doe
,
C. Q.
,
Chu-LaGraff
,
Q.
,
Wright
,
D. M.
and
Scott
,
M. P.
(
1991
).
The prospero gene specifies cell fates in the Drosophila central nervous system
.
Cell
65
,
451
464
.
FlyBase
(
1994
).
The Drosophila Genetic Database
.
Nucleic Acids Res
.
22
,
3456
3458
.
Fielding
,
C. J.
(
1967
).
Developmental genetics of the mutant grandchildless of Drosophila subobscura
.
J. Embryol. Exp. Morph
.
17
,
375
384
.
Goldstein
,
S.
(
1990
).
Replicative senescence: the human fibroblast comes of age
.
Science
249
,
1129
1133
.
Golic
,
K.
and
Lindquist
,
S.
(
1989
).
The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome
.
Cell
59
,
499
509
.
Gönczy
,
P.
,
Viswanathan
,
S.
and
DiNardo
,
S.
(
1992
).
Probing spermatogenesis in Drosophila with P-element enhancer detectors
.
Development
114
,
89
98
.
Goode
,
S.
,
Wright
,
D.
and
Mahowald
,
A. P.
(
1992
).
The neurogenic locus brainiac cooperates with the Drosophila EGF receptor to establish the ovarian follicle and to determine its dorsal-ventral polarity
.
Development
116
,
177
192
.
Harrison
,
D.
and
Perrimon
,
N.
(
1993
).
A simple and efficient generation of marked clones in Drosophila
.
Curr. Biol
.
3
,
424
433
.
Heberlein
,
U.
,
Wolff
,
T.
and
Rubin
,
G. M.
(
1993
).
The TGF-beta homolog dpp and the segment polarity gene hedgehog are required for propagation of a morphogenetic wave in the Drosophila retina
.
Cell
75
,
913
926
.
Herman
,
M. A.
and
Horvitz
,
H. R.
(
1994
).
The Caenorhabditis elegans gene lin-44 controls the polarity of asymmetric cell divisions
.
Development
120
,
1035
1047
.
Horvitz
,
H. R.
(
1988
).
Genetic control of Caenorhabditis elegans cell lineage
.
Harvey Lect
.
84
,
65
74
.
Karpen
,
G. H.
and
Spradling
,
A. C.
(
1992
).
Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp1187 by single P element insertional mutagenesis
.
Genetics
132
,
737
753
.
Kimble
,
J.
,
Crittenden
,
S.
,
Lambie
,
E.
,
Dodoyianni
,
V.
,
Mango
,
S.
and
Troemel
,
E.
(
1992
).
Regulation of induction by GLP1, a localized cell surface receptor in Caenorhabditis elegans
.
Cold Spring Harb. Symp. Quant. Biol
.
57
,
401
407
.
King
,
R. C.
(
1970
).
Ovarian Development in Drosophila melanogaster
.
New York
:
Academic Press
.
Koch
,
E. A.
and
King
,
R. C.
(
1966
).
The origin and early differentiation of the egg chamber of Drosophila melanogaster
.
J. Morph
.
119
,
283
304
.
Korden
,
E. C.
,
McKnight
,
R. A.
,
Jhappan
,
C.
,
Hennighausen
,
L.
,
Merlino
,
G.
and
Smith
,
G. H.
(
1995
).
Ectopic TGF-beta expression in the secretory mammary epithelium induces early senescence of the epithelial stem cell population
.
Dev. Biol
.
168
,
47
61
.
Lehmann
,
R.
and
Nusslein-Volhard
,
C.
(
1986
).
Abdominal segmentation, pole cell formation, and embryonic polarity require the localized activity of oskar, a maternal gene in Drosophila
.
Cell
47
,
141
152
.
Lin
,
H.
and
Spradling
,
A. C.
(
1993
).
Germ line stem cell division and egg chamber development in transplanted Drosophila germaria
.
Dev. Biol
.
159
,
140
152
.
Lin
,
H.
,
Yue
,
L.
and
Spradling
,
A. C.
(
1994
).
The Drosophila fusome, a germ line-specific organelle, contains membrane skeletal proteins and functions in cyst formation
.
Development
120
,
947
956
.
Lin
,
H.
and
Spradling
,
A. C.
(
1995
).
Fusome asymmetry and oocyte determination during Drosophila oogenesis
.
Dev. Genet
.
16
,
6
12
.
Lindsley
,
D.
and
Tokayasu
,
K. T.
(
1980
).
Spermatogenesis
.
In Genetics and Biology of Drosophila
(ed.
M.
Ashburner
and
T. R. F.
Wright
), pp.
225
294
.
New York
:
Academic Press
.
Mahowald
,
A. P.
and
Kambysellis
,
M. P.
(
1980
).
Oogenesis
.
In Genetics and Biology of Drosophila
(ed.
M.
Ashburner
and
T. R. F.
Wright
), pp.
141
224
.
New York
:
Academic Press
.
Matthews
,
K. A.
,
Miller
,
D. F. B.
and
Kaufman
,
T. C.
(
1989
).
Developmental distribution of RNA and protein products of the Drosophila α-Tubulin gene family
.
Dev. Biol
.
132
,
45
61
.
Metcalf
,
D.
(
1989
).
The molecular control of cell division, differentiation, commitment and maturation in haemopoietic cells
.
Nature
339
,
27
30
.
Patel
,
N. H.
(
1994
).
Imaging neuronal subsets and other cell types in whole mount Drosophila embryos and larvae using antibody probes
.
In Methods in Cell Biology
, Vol.
44
(eds.
L. S. B.
Goldstein
and
E.
Fyrberg
), pp.
445
487
.
New York
:
Academic Press
.
Patel
,
N. H.
,
Snow
,
P. M.
and
Goodman
,
C. S.
(
1987
).
Characterization and cloning of fasciclin III: a glycoprotein expressed on a subset of neurons and axon pathways in Drosophila
.
Cell
48
,
975
988
.
Pauli
,
D.
and
Mahowald
,
A. P.
(
1990
).
Germ line sex determination in Drosophila
.
Trends Genet
.
6
,
259
264
.
Podolsky
,
D. K.
(
1993
).
Regulation of intestinal epithelial proliferation: a few answers, many questions
.
Amer. J. Physiol
.
164
,
G179
186
.
Potten
,
C.
and
Loeffler
,
M.
(
1990
).
Stem cells: attribute, cycles, spirals, pitfalls and uncertainties. Lessons for and from the crypt
.
Development
110
,
1001
1020
.
Roholl
,
P. J.
,
Blauw
,
E.
,
Zurcher
,
C.
,
Dormans
,
J. A.
and
Theuns
,
H. M.
(
1994
).
Evidence for a diminished maturation of preosteoblasts into osteoblasts during aging in rats: an ultrastructural analysis
.
J. Bone Mineral Res
.
9
,
355
366
.
Rhyu
,
M. S.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1994
).
Assymmetric distribution of numb protein during division of the sensory organ precursor cell confers distinct fates to daughter cells
.
Cell
76
,
477
491
.
Ruohola
,
H.
,
Bremer
,
K. A.
,
Baker
,
D.
,
Swedlow
,
J. R.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1991
).
Role of neurogenic genes in establishment of follicle cell fate and oocyte polarity during oogenesis in Drosophila
.
Cell
66
,
433
449
.
Schüpbach
,
T.
,
Wieschaus
,
E.
, and
Nöthiger
,
R.
(
1978
).
A study of the female germ line in mosaics of Drosophila
.
Roux’s Arch. Dev. Biol
.
184
,
4156
.
Spradling
,
A. C.
(
1993
).
Developmental genetics of oogenesis
.
In Development of Drosophila melanogaster
. (ed.
Bate
,
M.
and
Martinez-Arias
,
A.
), pp.
1
70
.
Cold Spring Harbor, New York
:
Cold Spring Harbor Press
.
Uemura
,
T.
,
Shepherd
,
S.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1989
).
numb, a gene required in determination of cell fate during sensory organ formation in Drosophila embryos
.
Cell
58
,
349
360
.
Van Zant
,
G.
,
Holland
,
B. P.
,
Eldridge
,
P. W.
and
Chen
,
J.-J.
(
1990
).
Genotype-restricted growth and aging patterns of hematopoietic stem cell populations of allophenic mice
.
J. Exp. Med
.
171
,
1547
1565
.
Wieschaus
,
E.
,
Audit
,
C.
and
Masson
,
M.
(
1981
).
A clonal analysis of the roles of somatic cells and germ line during oogenesis in Drosophila
.
Dev. Biol
.
88
,
92
103
.
Wieschaus
,
E.
and
Szabad
,
J.
(
1979
).
The development and function of the female germ line in Drosophila melanogaster: A cell lineage study
.
Dev. Biol
.
68
,
29
46
.