The Division of Drosophila Germline Stem Cellsand Their Precursors Requires a Specific Cyclin
cystocytes in 16-cell cysts, nurse cells, and oocytes(Figures 1F and 1G). In GSCs, the CycB level reachesits peak in late G2, starts to decrease at metaphase,
Zhong Wang and Haifan Lin*Department of Cell Biology, BOX 3709Duke University Medical CenterDurham, North Carolina 27710 and is reduced to background levels by early anaphase
(see Figure S1 in the Supplemental Data available withthis article online). This expression profile suggests that,as expected, CycB accumulates in G2 phase to promoteSummarythe G2/M transition and is degraded in anaphase topromote exit from mitosis. In somatic cells, CycB ex-A fundamental yet essentially unexplored question inpression is also restricted to mitotic cells. It is presentstem cell biology is whether the stem cell cycle hasin somatic stem cells, prefollicular cells, and follicle cellsspecific features. Three B-cyclins in Drosophila,but is absent from postmitotic cells such as terminalCyclins (Cyc) A, B, and B3, associate with CDK1 andfilament, cap, and inner sheath cells (Figures 1F andplay partially redundant roles in embryogenic mitosis1G). Thus, CycB is expressed in all mitotic cells in larval[1, 2]. Here, we show that the division of Drosophilaand adult ovaries, suggesting its potential role in theGSCs and their precursors, the primordial germ cellsmitoses of these cells.(PGCs), specifically requires CycB. CycB is ubiqui-
tously expressed in both germline and somatic lin-eages. However, CycB mutation does not have obvi- CycB Is Required for PGC Proliferationous effect on somatic development but causes PGCs in the Developing Ovaryto severely under proliferate. Moreover, both female To study the potential requirement of CycB for PGCand male CycB mutant GSCs fail to be maintained proliferation, we examined third instar larval ovariesproperly. Removing Cyclin B specifically from female from CycB2, a null CycB mutant [2]. CycB2 mutant ovariesGSCs causes the same defect, confirming the direct have an average diameter of 95.5 � 13.8 nm (n � 6),and cell-autonomous function of Cyclin B for GSC divi- smaller than their heterozygous siblings (118.7 � 12.8sion. In contrast, two other G2 cyclins, CycA and nm, n � 10). Furthermore, a CycB2 ovary contains onlyCycB3, are also expressed in PGCs and GSCs, but 39 � 28 (n � 10) PGCs, as compared to 152 � 38 (n �overexpressing CycA cannot rescue the CycB mutant 10) in a wild-type ovary (Figure 2C), even though thedefects. These results indicate that the requirement mutant PGCs are morphologically normal (Figures 2Aof CycB for PGC and GSC divisions unlikely reflects and 2B). These data suggest that the proliferation ofthe insufficient level of G2 cyclins in the CycB mutant PGCs is severely affected without CycB. To determinebut is in favor of a distinct function of CycB in these the cell cycle defect that underlies PGC hypoprolifera-cells. Our results indicate that stem cells may use tion, we stained CycB2 mutant ovaries for a mitoticspecific cell cycle regulators for their division. marker, phosphorylated histone-3 (PH3) [4]. In wild-type
ovaries, all phases of mitosis are readily identified (Fig-ures 2D and 2F). However, in CycB2 mutant ovaries,Results and Discussionmost PGCs are arrested at late G2 phase, as indicated bytheir nuclear morphology and the PH3 staining patternCycB Is Expressed in PGCs and GSCs(Figures 2E and 2G). This indicates that CycB is requiredIt has been reported that CycB mutant flies have rudi-for PGC proliferation at least in part by promoting PGCmentary gonads [2]. To investigate whether CycB is in-entry into the M phase.volved in GSC division, we first examined its expression
during GSC development. GSCs are established at thelarval/pupal transition from the subset of PGCs that con- CycB Is Required for GSC Maintenancetact the somatic cap cells in the ovary [3] (Figure 1A). in the OvaryIn third instar larval ovaries, CycB protein is present at To follow the fate of PGCs in the CycB mutant duringvarious levels in PGCs, presumably reflecting the subsequent oogenesis, we examined the germline ofasynchrony of the PGC cell cycle (Figures 1C–1E). CycB the newly eclosed CycB2 mutant females. 47% of theis also present in all somatic cells of the gonad except mutant ovarioles are completely devoid of germ cellsfor postmitotic cells such as the forming terminal fila- (Figure S2B), whereas 35% of mutant ovarioles containment and cap cells (Figures 1C–1E). only a few germ cells (Figure S2C). The remaining 18%
In the adult ovary, a GSC divides asymmetrically to contain a few differentiating germline cysts and occa-self renew and to produce a differentiated daughter cell sionally a few egg chambers (Figures S2D and S2E).called a cystoblast, which undergoes four more rounds These results clearly indicate that none of the PGCs inof incomplete mitoses to produce a 16-cell cyst (Figure the CycB mutant larval ovary becomes self-renewing1B). CycB is expressed at high levels in a subset of GSCs in the adult ovary. In addition, CycB is involvedGSCs, cystoblasts, and 2- to 8-cell cysts (Figure 1F) in cystoblast and/or cystocyte division because CycBbut is again absent from postmitotic cells, including mutant egg chambers often contain fewer than 16 cells
(Figure S2E).The failure of PGCs to become self-renewing GSCs*Correspondence: [email protected]
Cyclin B in Germline Stem Cell Division329
Figure 1. CycB Expression during PGC De-velopment and Early Oogenesis
(A and B) Schematic drawings of a third instarlarval ovary (A) and an adult germarium (B).In the larval ovary (A), nascent cap cells (CpC)are located at the base of terminal filaments(TF). PGCs that contact cap cells (GSC*) arefated as GSCs. In the germarium (B), 2–3GSCs, located in the apical region in contactwith cap cells, divide asymmetrically to pro-duce a daughter GSC and a cystoblast (CB)that divides four more times with incompletecytokinesis to form a 16-cell cyst. Spectro-somes (Sp) mark GSCs and cystoblasts,whereas fusomes (Fu) mark 2-, 4-, 8-, and 16-cell cysts. Within the 16-cell cyst, 15 cellslater differentiate into nurse cells (NC), andone becomes oocyte (O). Somatic stem cells(SSC) are separated from cap cells by innersheath cells (IS). SSCs produce prefollicular(PFC) and follicle cells (FC) that envelop 16-cell cysts to form egg chambers.(C–E) Confocal images of a third instar ovarystained for CycB (C, red) and Vasa to markPGCs (D, green) with the images merged in(E). (F) A germarium stained for CycB (red)and spectrosomes/fusomes (green). (G) Astage 2 egg chamber stained for CycB (red),PH3 (green), and DAPI (blue).
could reflect either a direct requirement for CycB in tants (Figure 3D). At day 8 post induction, a wild-typeGSC produces 2.7 � 0.8 (n � 40) germline cysts inGSCs during oogenesis or an indirect consequence of
PGC defects prior to oogenesis. In addition, it could be the germarium, whereas a CycB2 mutant GSC produces2.0 � 0.8 (n � 26) cysts. By day 14, a wild-type GSCa side effect of an ovarian somatic defect in the CycB
mutants. To distinguish between these possibilities, produces 3.5 � 1.0 (n � 50) cysts in the ovariole, whereasa mutant GSC produces only 1.1 � 0.4 (n � 8) cysts.we removed CycB specifically from GSCs by inducing
CycB2 mutant germline clones in the adult ovary by FLP-FRT-mediated mitotic recombination (see ExperimentalProcedures). Clones were identified by their lack of ex- CycB Is Required for GSC Maintenance
in the Testispression of the armadillo-LacZ reporter gene. The self-renewing ability of GSCs was measured with two criteria: Because CycB mutant males are also sterile [2], we
investigated whether CycB is required for male GSC(1) change in the percentage of GSC clones in an ovary atvarious time points after clone induction, which directly division. Like oogenesis, Drosophila spermatogenesis
begins with asymmetric GSC divisions that each pro-measures the rate of GSC loss; and (2) clone size (i.e.,the number of cysts and egg chambers derived from a duces a daughter GSC and a differentiating daughter
cell (for review, see [6]). The latter, called a gonialblast,clonal GSC), which directly measures how many timesa clonal GSC has divided. Because it normally takes 7 further divides four times to produce a germline cyst
containing 16 interconnected primary spermatocytes.days for a GSC to develop into an egg chamber [5],clones present in the germarium 8 days after induction The spermatocytes then grow in size and enter meiosis.
Because male GSCs start self-renewing divisions at theshould be derived from GSCs. Wild-type GSC cloneswere observed in 21% � 6% (n � 188) of the ovarioles first instar larval stage (reviewed in [7]), we examined
the defect of third instar larval and young adult CycBat day 8 after clonal induction, whereas CycB2 mutantGSC clones were present only in 11% � 0.6% (n � 231) mutant testes by staining with the anti-VASA antibody
for germ cells and anti-1B1 antibody for spectrosomesof ovarioles, 52% of the wild-type level (Figures 3A–3C).At 14 days after clone induction, although wild-type GSC and fusomes. At this time, a wild-type testis contains
not only GSCs but also many differentiating germlineclones were present in 34% � 6.8% (n � 155) of theovarioles, CycB2 mutant GSCs were present in 7% � cysts up to the spermatocyte stage (Figure 4A). In con-
trast, the CycB2 mutant testis contains few or no GSCs4.2% (n � 118) of the ovarioles, only 20% of the wild-typelevel. This increasing difference after clone induction and only a small number of germline cysts (Figure 4B).
By the adult stage, the wild-type adult testis contains aindicates that CycB mutant GSCs are progressively lostduring oogenesis. full complement of germ cells (Figure 4C), whereas the
adult CycB2 mutant testis contains only a few germ cellsThe above conclusion is further supported by measur-ing the clone size of GSCs in wild-type and CycB mu- (Figure 4D). These results indicate that CycB is required
Current Biology330
Figure 2. CycB Mutant PGCs Arrest at the G2 Phase in Larval Ovaries
(A and B) Confocal images of a wild-type and a CycB2 mutant third instar larval ovary, respectively, stained with anti-Vasa (green) and anti-1B1 (red) antibodies.(C) The number of PGCs in wild-type (WT) and CycB mutant ovaries (ovary, n � 10).(D–G) Wild-type (D and F) and CycB2 mutant (E and G) third instar larval ovaries stained for PH3. Germline cells are distinguished from somaticcells by their large and round nuclei (nucleus staining not shown). TF region is outlined in (D) and (E). In wild-type ovaries (D), mitotic PGCsare readily detected at prometaphase (pm), anaphase (a), and telophase (t), whereas in CycB2 mutants (E), most of PGCs are arrested at G2phase with prominent nuclei displaying punctate PH3 staining (arrows). (F) and (G) are high-magnification views of a prometaphase PGC inwild-type (F) and a G2-arrested PGC in CycB2 mutant (G) ovaries. The bar in (A) is for (A), (B), (D), and (E); the bar in (F) is for (F) and (G).
Figure 3. CycB Is Required Cell Autonomously for Female GSC Proliferation
(A and B) Wild-type (A) and CycB2 mutant (B) GSC clones generated in the adult germaria (see Supplemental Data). LacZ (red)-negative cellsare clones derived from a single GSC. The anti-Vasa antibody (green) marks all germ cells.(C) The clone induction efficiency (the percentage of clone-containing ovarioles) of wild-type (WT) and CycB2 mutant GSCs was measured at8 and 14 days after clone induction.(D) The number of germline cysts derived from a single wt or CycB2 mutant GSC in an ovariole.
Cyclin B in Germline Stem Cell Division331
Figure 4. CycB Is Required for Male GSCSelf-Renewing Division
Confocal images of wild-type (A and C) andCycB2mutant (B and D) third instar larval (Aand B) and adult (B and D) testes stained forVasa (green) and 1B1 (red). Similar to femaleGSCs and cystoblasts, male GSCs and gonial-blasts are marked by spectrosomes, whereasearly germline cysts are marked by fusomes.Sp cyst, spermagonia cysts; Spc cyst, sperm-atocyte cyst.
for the self-renewing division of male GSCs and for the cells. Both CycA and CycB3 are coexpressed with CycBin PGCs and GSCs. In these cells, CycA is localized tofurther development of testicular germline cysts.the cytoplasm, nucleus, and the fusomes at differentphases of the cell cycle, whereas CycB3 is predomi-CycB Is Dispensable for Somatic Cell Proliferation
In contrast to its germline requirement, CycB appears to nantly nuclear [9] (Figure S3 and data not shown). Fur-thermore, both CycA and CycB3 are expressed in CycB2be dispensable for somatic cell divisions in both sexes.
First, the proliferation of somatic cells in CycB mutant mutant PGCs and GSCs (data not shown). This rulesout the possibility that the requirement of CycB is duelarval ovaries is largely normal. There are comparable
number of terminal filaments formed in a wild-type (18.3 � to its unique expression or subcellular localization. Thisconclusion is further supported by a recent demonstra-1.5, n � 10) and CycB2 mutant third instar ovary of similar
age (17.8 � 1.5, n � 10), with on average 11–13 cells tion that the function of CycA does not depend on itssubcellular localization [10].for each terminal filament. Both apical and basal somatic
cells display normal morphology and organization (cf. To determine whether the requirement for CycB re-flects an insufficient dose of G2 cyclins in the CycBFigures 2A and 2B). Second, somatic cells also prolifer-
ate normally in CycB2 mutant adult ovary, producing mutants, we overexpressed CycA in the CycB mutant byheat shocking the CycB2 mutant containing a transgenicnormal number of follicle cells that envelop germline
cysts (Figures S2B–S2E). The frequency and size of a hs-CycA gene (hs, hsp70 promoter). Heat shock-induced expression of this transgene has been reportedCycB mutant follicle cell clone are comparable to that
of a sister clone (data not shown), indicating the division to be able to rescue the cell cycle block in CycA;CycBdouble mutant embryos [1] and also produce 32-cellof neither somatic stem cell nor follicle cell is severely
affected. Third, testicular somatic cells, including the phenotype in wild-type adult ovaries [9]. We verified thefunction of hs-CycA by heat shock-induced overexpres-hub cells, appear to be normal in both third instar and
adult mutant testes (Figures 4B and 4D), suggesting sion of this gene in adult females, which reproduced the32-cell germline cyst phenotype. We then determinedthat CycB is dispensable for somatic cells in the testis.
Finally, the null CycB mutant flies are viable, even though the ability of this hs-CycA gene to rescue the CycBmutant phenotype with hs-CycB gene as a positive con-the mitoses in CycB mutant embryos are not entirely
normal [1]. This further indicates that CycB is not re- trol. CycA was overexpressed in the CycB2 mutant atdifferent stages of germline development to see if thisquired in the extraovarian somatic lineages.will rescue the CycB mutant defect in PGC and GSCproliferation. Heat shocks starting at second and thirdCycA Cannot Functionally Substitute CycB
in the Germline instar larval stage allowed hs-cycB to rescue the CycBmutant phenotype to almost completion (93% and 80%,Why is CycB ubiquitously expressed in both somatic
and germline cells yet specifically required for the GSC respectively; Figure S4). Even heat shocks starting fromnewly eclosed stage to 4-day-old adult showed 8% oflineage? The currently favored theory is that the func-
tional differences between the same types of cyclins rescue. Under none of these conditions, hs-CycA evenpartially rescued the CycB mutant phenotype (Figureare largely determined by when and where they are
expressed, rather than by which substrates they induce S4). These analyses indicate that the requirement ofCycB for PGC and GSC divisions unlikely reflects theCDKs to phosphorylate [8]. To determine whether CycB
is the only G2 cyclin expressed in PGCs and GSCs, we insufficient level of G2 cyclins in the CycB mutant butis in favor of a distinct function of CycB in these cells.examined the expression of CycA and CycB3 in these
Current Biology332
spectrosome was counted as one PGC. Total PGC number in anThis conclusion is supported by the previous report thatovary was determined by manually marking every nonoverlappingCycB3 is not required for PGC or GSC divisions [2].PGC from each slice. Ten ovaries from both wild-type and CycB2
mutants were counted this way.ConclusionsHere, we have demonstrated that a specific cyclin is
Supplemental Datarequired for the division of the GSC lineage in Drosoph- Supplemental Data include four figures and can be found with thisila. Our results suggest that stem cells and/or the germ- article online at http://www.current-biology.com/cgi/content/full/
15/4/328/DC1/.line may require distinct cell cycle components for theirdivisions. It has been reported that maternal CycBmRNA is enriched but translationally repressed in mi- Acknowledgmentsgrating embryonic PGCs [11, 12]. However, the function
We thank Xiao Shi for PGC quantification and image in Figure 2Bof the zygotic CycB remains elusive. This study demon-and Brian Freibaum for images in Figures 2A, 4A, and 4B. We alsostrates a requirement of zygotic CycB in the PGCs andthank Drs. Brigid Hogan and Seth Findley for their valuable com-
GSCs. In addition, our results indicate that such a re- ments and Dr. Yue Xiong for discussions. This work was supportedquirement may reflect a distinct function of CycB in by the National Institutes of Health (HD33760).these cells. It is interesting to speculate that these fea-tures of cell cycle regulation in the Drosophila GSC lin- Received: October 1, 2004eage may exist in other stem cell systems as well. Revised: December 8, 2004
Accpeted: December 10, 2004Experimental Procedures Published: February 22, 2005
Fly Stocks and Heat Shock ReferencesAll Drosophila stocks were grown at 25�C on yeast-containing cornmeal/molasses medium. CycB2, a null mutant previously described 1. Knoblich, J.A., and Lehner, C.F. (1993). Synergistic action ofin [2], was a gift from Dr. Chris Lehner. The hsFlp12;FRT42D arma- Drosophila cyclins A and B during the G2-M transition. EMBOdillo-lacZ flies were described in [13]. Hs-CycB and hs-CycA trans- J. 12, 65–74.genic flies were described in [1, 14] and were gifts from Drs. Pat 2. Jacobs, H.W., Knoblich, J.A., and Lehner, C.F. (1998). Drosoph-O’Farrel and Chris Lehner. ila Cyclin B3 is required for female fertility and is dispensable
for mitosis like Cyclin B. Genes Dev. 12, 3741–3751.Germline Clonal Analysis 3. Zhu, C.H., and Xie, T. (2003). Clonal expansion of ovarian germ-To study the phenotype of CycB mutant clones, we utilized FLP- line stem cells during niche formation in Drosophila. Develop-mediated recombination [13, 15, 16]. CycB2 was recombined onto ment 130, 2579–2588.the FRT42D chromosome. The presence of CycB2 mutation was 4. Hendzel, M.J., Wei, Y., Mancini, M.A., Van Hooser, A., Ranalli,verified by its homozygous sterility, whereas the FRT site was de- T., Brinkley, B.R., Bazett-Jones, D.P., and Allis, C.D. (1997).tected by PCR. To generate germline clones, w;FRT42D CycB2/CyO Mitosis-specific phosphorylation of histone H3 initiates primar-males were crossed to virgin females of w hsFlp12;FRT42D armLacZ, ily within pericentromeric heterochromatin during G2 andand newly eclosed progeny lacking the CyO chromosome were heat spreads in an ordered fashion coincident with mitotic chromo-shocked at 37�C for 30 min twice daily with 11.5 hr intervals. Eight some condensation. Chromosoma 106, 348–360.days after the final heat shock, ovaries of w hsFlp12;FRT42D CycB2/ 5. Wieschaus, E., Audit, C., and Masson, M. (1981). A clonal analy-FRT42D armLacZ females were dissected and stained with the anti- sis of the roles of somatic cells and germ line during oogenesis�-Gal antibody to identify clones. Anti-VASA antibody and DAPI in Drosophila. Dev. Biol. 88, 92–103.were used to visualize germ cells and nuclei, respectively. Clones 6. Fuller, M.T. (1998). Genetic control of cell proliferation and differ-were identified by their lack of the expression of the armadillo-LacZ entiation in Drosophila spermatogenesis. Semin. Cell Dev. Biol.reporter gene. 9, 433–444.
7. Fuller, M.T. (1993). Spermatogenesis. In The Development ofImmunofluorescence Staining and Microscopy Drosophila Melanogaster, Volume I, M. Bate and A.M. Arias,Drosophila ovaries were dissected and stained according to [17]. eds. (New York: Cold Spring Harbor Laboratory Press), p. 103.The rabbit anti-VASA antibody [18], a gift from Dr. Yuh-Nung Jan, 8. Murray, A.W. (2004). Recycling the cell cycle: cyclins revisited.was used at 1:1000 dilution to label germline cells. The rabbit anti- Cell 116, 221–234.PH3 antibody (Upstate Biotech, Lake Placid, NY) was used at 1:1000 9. Lilly, M.A., de Cuevas, M., and Spradling, A.C. (2000). Cyclin Adilution. Monoclonal anti-CycB (F2F4; used at 1:100 dilution) [1], associates with the fusome during germline cyst formation inanti-CycA (used at 1:5 dilution) [1], and anti-1B1 [19] (used at 1:2 the Drosophila ovary. Dev. Biol. 218, 53–63.dilution) antibodies were obtained from the University of Iowa Hy- 10. Dienemann, A., and Sprenger, F. (2004). Requirements of cyclinbridoma Bank. Monoclonal anti-�-Gal antibody (Promega, Inc., A for mitosis are independent of its subcellular localization.Madison, WI) was used at 1:300 dilution. Rabbit anti-CycB3 antibody Curr. Biol. 14, 1117–1123.was a gift from Dr. C. Lehner [2] and was used at 1:600 dilution. 11. Dalby, B., and Glover, D.M. (1992). 3� non-translated sequencesFluorescently labeled secondary antibodies (Jackson Immunore- in Drosophila cyclin B transcripts direct posterior pole accumu-search Laboratories, West Grove, PA) were used at 1:200 dilution lation late in oogenesis and peri-nuclear association in syncytialfrom a 1 mg/ml stock. The DNA dye DAPI (SIGMA, St. Louis, MO) embryos. Development 115, 989–997.was used at 0.5 �g/ml. 12. Asaoka-Taguchi, M., Yamada, M., Nakamura, A., Hanyu, K., and
All samples were examined by epifluorescence microscopy by Kobayashi, S. (1999). Maternal Pumilio acts together with Nanosusing a Zeiss Axioplan microscope with images captured by a Pho- in germline development in Drosophila embryos. Nat. Cell Biol.tometrics CoolSnap HQ CCD camera (Phoenix, AZ) with IP Lab 1, 431–437.software. Confocal microscopy was done with a Zeiss LSM510 sys- 13. Lecuit, T., and Cohen, S.M. (1997). Proximal-distal axis forma-tem. All images were processed with Adobe Photoshop 6.0. tion in the Drosophila leg. Nature 388, 139–145.
14. Sigrist, S., Jacobs, H., Stratmann, R., and Lehner, C.F. (1995).Exit from mitosis is regulated by Drosophila fizzy and the se-Quantifying PGC Number in Third Instar Larval Ovaries
1 �m optical slices that span a whole third instar larval ovary were quential destruction of cyclins A, B and B3. EMBO J. 14, 4827–4838.obtained from either wild-type or CycB2 mutants with the Zeiss
LSM510 system. A single cell with both VASA staining and a single 15. Chou, T.B., and Perrimon, N. (1996). The autosomal FLP-DFS
Cyclin B in Germline Stem Cell Division333
technique for generating germline mosaics in Drosophila mela-nogaster. Genetics 144, 1673–1679.
16. Chou, T.B., and Perrimon, N. (1992). Use of a yeast site-specificrecombinase to produce female germline chimeras in Drosoph-ila. Genetics 131, 643–653.
17. Lin, H., Yue, L., and Spradling, A.C. (1994). The Drosophila fu-some, a germline-specific organelle, contains membrane skele-tal proteins and functions in cyst formation. Development 120,947–956.
18. Hay, B., Jan, L.Y., and Jan, Y.N. (1990). Localization of vasa,a component of Drosophila polar granules, in maternal-effectmutants that alter embryonic anteroposterior polarity. Develop-ment 109, 425–433.
19. Zaccai, M., and Lipshitz, H.D. (1996). Differential distributionsof two adducin-like protein isoforms in the Drosophila ovaryand early embryo. Zygote 4, 159–166.
