l(3)malignant brain tumor and Three Novel Genes Are Required for Drosophila Germ-Cell Formation

Corresponding author: Ruth Lehmann, Skirball Institute, 4th Floor, 540 First Ave., New York University School of Medicine, New York, NY 10016., lehmann{at}saturn.med.nyu.edu (E-mail)

Communicating editor: T. SCHUPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To identify genes involved in the process of germ-cell formation in Drosophila, a maternal-effect screen using the FLP/FRT-ovo^D method was performed on chromosome 3R. In addition to expected mutations in the germ-cell determinant oskar and in other genes known to be involved in the process, several novel mutations caused defects in germ-cell formation. Mutations in any of three genes [l(3)malignant brain tumor, shackleton, and out of sync] affect the synchronous mitotic divisions and nuclear migration of the early embryo. The defects in nuclear migration or mitotic synchrony result in a reduction in germ-cell formation. Mutations in another gene identified in this screen, bebra, do not cause mitotic defects, but appear to act upstream of the localization of oskar. Analysis of our mutants demonstrates that two unique and independent processes must occur to form germ cells—germ-plasm formation and nuclear division/migration.

CELL type specification is a crucial process in the development of multicellular organisms. In many organisms, one of the first cell types specified in the developing embryo are the primordial germ cells (PGCs). These cells eventually develop as sperm and egg. One key step in the specification of PGCs in Drosophila is the establishment of a specialized cytoplasm at the posterior of the egg. This cytoplasm, called the pole plasm or germ plasm, consists of maternally supplied RNAs and proteins that are deposited and localized during oogenesis. A crucial component of the establishment of germ plasm is oskar—the osk mRNA is localized to the posterior and translated there, and OSK protein organizes and recruits the remaining germ-plasm components (EPHRUSSI et al. 1991 ; KIM-HA et al. 1991 ). osk is both necessary and sufficient for germ-plasm assembly, as mislocalization of osk to the anterior of the embryo leads to ectopic germ-cell formation there (EPHRUSSI and LEHMANN 1992 ). The germ plasm also has a second function—abdomen formation—which is accomplished through the recruitment of nanos (nos) RNA to the posterior by OSK.

PGCs form prior to any of the somatic cells of the embryo. In the early Drosophila embryo, the male and female pronuclei fuse and then undergo 13 rounds of synchronous mitoses without cell division to produce a syncytium (ZALOKAR and ERK 1976 ; FOE and ALBERTS 1983 ). Only after the fourteenth cycle does somatic cellularization occur. Germ cells, on the other hand, are formed at an earlier stage. During cycles 6–8, the dividing nuclei undergo an actin and myosin-dependent axial expansion in which they spread out internally along the A-P axis (ZALOKAR and ERK 1976 ; HATANAKA and OKADA 1991 ; WHEATLEY et al. 1995 ; ROYOU et al. 2002 ). During cycles 8 and 9, the nuclei migrate in a microtubule-dependent process to the cortex of the syncytial embryo (ZALOKAR and ERK 1976 ; BAKER et al. 1993 ). At the posterior, pole buds, the precursors to cellularized germ cells, are forming. At cycle 10, the posterior-most nuclei, along with the germ plasm surrounding them, are incorporated into the PGCs as the pole buds cellularize. These cells then break from the cycling synchrony of the rest of the nuclei, undergoing two to three additional, asynchronous rounds of cell division. The remaining nuclei continue their migration to the periphery of the embryo, completing the last four rounds of synchronous mitoses until somatic cellularization (ZALOKAR and ERK 1976 ; FOE and ALBERTS 1983 ).

Prior to the eighth mitotic cycle, the zygotic genome is transcriptionally silent (PRITCHARD and SCHUBIGER 1996 ). During this time of transcriptional quiescence (and for some time afterward), the embryo depends on maternally provided proteins and RNAs. For example, the synchronous mitotic divisions are regulated by maternal stores of cyclins A, B, and B3 (EDGAR et al. 1994 ; JACOBS et al. 1998 ; STIFFLER et al. 1999 ). Also, initial patterning is dependent upon nanos and bicoid RNAs (among others) from the mother (ST. JOHNSTON and NUSSLEIN-VOLHARD 1992 ). Those components of the germ plasm that are responsible for germ-cell formation are also maternal, especially given that transcriptional silence continues even longer in the newly formed PGCs then in the soma (VAN DOREN et al. 1998 ; SEYDOUX and DUNN 1997 ). Thus the maternal genotype is critical for PGC formation in the embryo.

Genetic screens have contributed much to what we know about how germ cells are formed in Drosophila. However, many of the genes implicated genetically in germ-cell formation were in fact discovered serendipitously as members of the posterior group genes. Genes in this group, osk among them, were initially identified on the basis of their posterior patterning defects and not on their defects in germ-cell formation (NUSSLEIN-VOLHARD et al. 1987 ). These germ-plasm mutations were selected for their defect in the second function of the germ plasm, posterior axis determination in the embryo through the function of nos.

Here, we describe a screen to identify defects in germ-cell formation specifically. As gene products involved in this process are expected to be contributed maternally, germ-line clones were made in otherwise heterozygous animals so that potentially lethal mutations could be screened. We used the yeast FLP recombinase and its recognition site (FRT) in combination with the dominant female sterile mutation ovo^D (CHOU and PERRIMON 1996 ) to make germ-line clones of EMS mutations on the right arm of chromosome 3. Several mutants that affect germ-cell formation were identified. One gene is involved in the formation of the germ plasm, while three other genes are required for nuclear migration and cycling.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks:
The following alleles were identified in the screen and used for phenotypic analysis: osk^GM52, l(3)mbt^GM76, l(3)mbt^GM79, l(3)mbt^GM161, shkl^GM45, shkl^GM130, shkl^GM163, oosy^GM73, oosy^GM47, beb^GM29, and beb^GM50. Df(3R)mbt^PE3 (WISMAR et al. 1995 ), l(3)mbt^E2 (LOFFLER 1988 ), and l(3)mbt^ts1 (GATEFF 1978 ) were obtained from Jasmine Wismar. Transgenic flies used to create starting stocks for the screen, osk¹⁶⁶ and Df(3R)D605, were obtained from the Bloomington Stock Center or were present in the laboratory. All flies were raised at 25° unless otherwise indicated.

Mutagenesis:
See Fig 1 for a schematic of the screen. A recently isogenized line containing FRT sequences near the centromere on chromosome 3R (82B) with e as a marker and the fat-facets-lacZ (P{faf-lacZ}) transgene (FISCHER-VIZE et al. 1992 ) on the X was used for the mutagenesis strain. This line was selected for its low frequency of germ-cell migration loss. A total of 8550 w P{faf-lacZ}; P{FRT}82B e males were mutagenized with 25–35 mM EMS (Sigma, St. Louis) in 1% sucrose for 16–24 hr as described (ASHBURNER 1989A ) with the modification that they were starved on a water-soaked Kimwipe for 6 hr prior to mutagenesis. These males were mated to 8600 virgin females of the genotype w P{faf-lacZ}; Pr Dr/TM3 Sb P{hs:hid}. The crosses were kept at room temperature (22°) and the males were discarded after 5 days to avoid clonal mutations. Approximately 21,600 single virgin females from the F₁ generation with the genotype w P{faf-lacZ}; P{FRT}82B e/TM3 Sb P{hs:hid} were each mated to two or three males of the genotype y w P{hs:flp}/Y ; P{FRT}82B P{ovo^D}/TM3 Sb P{hs:hid}. The hs:flp transgene (P{hs:flp}) allows for expression of the yeast FLP recombinase upon heat induction (GOLIC and LINDQUIST 1989 ; GOLIC 1991 ), while the ovo^D (P{ovo^D}) transgene is a dominant female-sterile mutation that allows for positive selection of female germ-line clones (CHOU and PERRIMON 1996 ). These crosses were allowed to lay eggs for 4 days, when the parents were discarded. The vials were subjected to heat shock on day 6 by placing the vials in a 37° water bath for 2 hr. Heat-shock induction of the FLP recombinase produced mutant clones in the otherwise heterozygous flies, while all flies containing the TM3 balancer were killed by virtue of the hs:hid transgene (P{hs:hid}) (GRETHER et al. 1995 ; MOORE et al. 1998 ). The F₂ progeny were transferred to a vial containing yeast for 2–3 days. Embryos from these crosses were collected and stained for ß-galactosidase activity from the germ-line-specific faf-lacZ transgene as described (MOORE et al. 1998 ). Balanced lines were established from the F₂ males.

View larger version (11K): In this window In a new window Download PPT slide   Figure 1. Crossing scheme to establish females heterozygous for a mutagenized third chromosome and the dominant female-sterile mutation ovoD1. Both third chromosomes contain FRT sequences near the centromere of the right arm. Heat shock as larvae induces the FLP recombinase and kills flies with the balancer. A faf-lacZ fusion gene is used as a marker for germ line. (For an explanation of stocks used, see MATERIALS AND METHODS). *, the mutagenized chromosome.

Complementation testing, mapping, and allele sequencing:
Mutant lines were tested for complementation by anti-Vasa immunostaining of embryos laid by trans-heterozygous females. Once complementation groups were established, representative lines were complementation tested against relevant deletion strains from the deficiency kit maintained by the Bloomington Stock Center or obtained from other laboratories. Candidate transposon insertions and mutant alleles were complementation tested as well. All deficiency and mutant lines used complemented the shkl, oosy, and beb alleles. For mapping, one allele of each complementation group was crossed to a ru st cu sr e+ ca mapping strain. Trans-heterozygous P{FRT}82B e allele¹/ru h st cu sr e+ ca females were crossed with ru h st cu sr Pr e ca/TM6 males. Single males carrying the recombinant chromosome ("Rec") in trans to ru h st cu sr Pr e ca were crossed to P{faf-lacZ}; P{FRT}82B e allele²/TM3 Sb females. In the next generation the P{faf-lacZ}; P{FRT}82B e allele²/Rec female progeny was tested for the mutant germ-cell phenotype by staining for ß-galactosidase activity from the PGC-specific faf-lacZ transgene. The map positions for shkl, oosy, and beb are listed in Table 1.

One complementation group with three members failed to complement deletion lines that uncover l(3)mbt. This group proved to be allelic to l(3)mbt. DNA from flies trans-heterozygous for each allele of l(3)mbt and a deletion of the l(3)mbt gene [Df(3R)D605] was prepared as described (ASHBURNER 1989B ). Multiple independent PCR reactions were used to amplify each exon of the gene prior to pooling for sequencing in a ABI Prism 3700 machine (Rockefeller University DNA Sequencing Resource Center). SeqMan II (DNASTAR, Madison, WI) and EditView (Applied Biosystems, Foster City, CA) were used for analysis of sequencing data.

Whole-mount immunostaining and in situs:
Embryos were fixed (after dechorionation in 50% bleach for 5 min) by gentle shaking for 20 min in 8 ml heptane, 0.25 ml 37% formaldehyde, 1.75 ml PBS, followed by devitellinization by addition of methanol. Following rehydration, primary antibody staining was carried out in 0.2% Tween in PBS overnight at 4°. Biotinylated secondary antibody (1:2000; Roche, Indianapolis) followed by Vectastain ABC kit (Vector Laboratories, Burlingame, CA) and 3,3'-diaminobenzidine (DAB) detection were performed as described (MOORE et al. 1998 ). Embryos were embedded in PolyBed812 (Polysciences, Niles, IL) and analyzed using a Zeiss Axiophot and a Sony digital camera with Adobe Photoshop software. Antibodies used were anti-Vasa at 1:5000 and anti-phospho-Histone (H3; Upstate Biotechnology, Lake Placid, NY) at 1:1000. For anti-Centrosomin (T. Kaufman) labeling, embryos were fixed by permeabilization in octane for 30 sec followed by fixation in methanol for 1–2 hr. After rehydration and primary antibody incubation (1:1000 overnight at 4°), a Cy3-conjugated secondary (Jackson ImmunoResearch Laboratories, West Grove, PA) was used at 1:500. DNA was stained with OliGreen (Molecular Probes, Eugene, OR) at 1:5000 with the addition of 5 µg/ml RNase A. Embryos were mounted in 50% glycerol, 2.5% DABCO, and PBS and analyzed using a Leica TCS/NT confocal microscope.

Whole-mount in situ hybridizations were performed as described (LEHMANN and TAUTZ 1994 ). Embryos were hybridized at 55° for 18 hr. Antisense digoxigenin-labeled RNA probes were synthesized with the Genius kit (Roche). The oskar (C. Rongo), gcm (T. Jongens), and pgc (S. Kobayashi) probes were made from pBluescript plasmids containing the respective cDNAs. The gcm probe was made from a pNB40-derived cDNA clone. Embryos were mounted and analyzed as described above for DAB labeling.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of mutants affecting germ-cell number:
To identify genes involved in the formation of PGCs in Drosophila, we undertook a screen for recessive maternal mutations. We used the FLP-FRT/ovo^D system (CHOU and PERRIMON 1996 ), in which clones were generated by the yeast FLP recombinase in animals heterozygous for EMS-induced mutations. Homozygous mutant clones generated in the germ line were selected by loss of a dominant female-sterile mutation (ovo^D). Embryos laid by these females were screened for phenotypes caused by mutation or loss of a maternally supplied gene product (see Fig 1 and MATERIALS AND METHODS). To identify defects in the germ cells, a transgene expressing a fusion of the Fat-facets protein and ß-galactosidase enzyme was used (FISCHER-VIZE et al. 1992 ). This protein localizes to the germ cells and is an excellent marker for germ cells throughout embryogenesis.

Approximately 10,000 mutagenized lines were screened. Slightly more than one-quarter of these lines (2699) did not lay any eggs, suggesting that the germ-line clones caused oogenesis defects. These lines were further characterized in a separate oogenesis screen (MORRIS et al. 2003 ). Eggs were collected from the remaining lines (7346) and stained for ß-galactosidase activity. Lines that had a change in the number of germ cells were kept and established as stable stocks.

Initially, 166 lines were maintained. These were rescreened using an antibody against the Vasa protein (which is maternally deposited and localized to the germ cells). Many lines either had an inconsistent germ-cell reduction phenotype or had significant patterning/developmental defects beyond the germ-cell phenotype. However, 14 lines had a consistent reduction in germ-cell number with minimal patterning or other defects and were therefore selected as specifically affecting germ-cell formation. These 14 lines were subjected to complementation analysis with each other, with genes on 3R known to affect germ-cell formation, and with the other lines from the screen that showed, as clones, inconsistent phenotypes or pleiotropic effects. Although some lines are homozygous lethal, all trans-heterozygous animals are viable. Lethal mutations could often be mapped to a second site via complementation with deficiencies. Alleles in three genes were previously known to affect specifically germ-cell formation and map to chromosome 3R. Five oskar alleles and one Tropomyosin I allele were identified. We did not identify new alleles of barentsz (LEHMANN and NUSSLEIN-VOLHARD 1986 ; ERDELYI et al. 1995 ; VAN EEDEN et al. 2001 ). This suggests that the scale of the screen approached saturation of 3R. Three alleles of a previously identified tumor suppressor gene, l(3)malignant brain tumor [l(3)mbt], were shown to have germ-cell formation defects. Additionally, six novel complementation groups were also identified. Three of these were single alleles and are not discussed further. The other complementation groups are shackleton (shkl) with three alleles and out of sync (oosy) and bebra (beb), both with two alleles (see Table 1 for alleles and map position).

The maternal-effect phenotypes of these mutations [l(3)mbt, shkl, oosy, and beb] are shown in Fig 2. Embryos laid by females trans-heterozygous for two mutant alleles of a given gene were examined. For the remainder of this work, these are referred to as mutant embryos. These mutant embryos were stained with an antibody to the Vasa protein. The number of germ cells was greatly reduced in all four mutants (Table 1). In some embryos, no germ cells were present. This reduction results from failure to initially form germ cells, as opposed to formation and subsequent death, since the defect is visible at early stages (Fig 2C, Fig E, Fig G, and Fig I) as well as at later stages of embryogenesis (Fig 2D, Fig F, Fig H, and Fig J). In these four mutants, the number of germ cells formed varied within a collection of embryos, in that a range (from none to almost wild-type numbers) could occur. This is unlike most osk alleles, where the lack of germ cells is consistent and penetrant (LEHMANN and NUSSLEIN-VOLHARD 1986 ). Additionally, many of the mutant embryos from l(3)mbt, shkl, oosy, and beb complete embryogenesis, hatch, and produce viable, fertile adults.

View larger version (79K): In this window In a new window Download PPT slide   Figure 2. Embryos laid by mutant females fail to form wild-type numbers of germ cells. Anterior is to the left. Germ cells are visualized with an anti-Vasa antibody. (A, C, E, G, and I) Stage 5 embryos. (B, D, F, H, and J) Stage 9–10 embryos. (A and B) wild type; (C and D) l(3)mbtGM79/Df; (E and F) shklGM45/GM163; (G and H) oosyGM47/GM73; (I and J) bebGM29/GM50. The number of germ cells formed is severely reduced (in some cases to zero) in embryos laid from all mutant lines (C–J).

Only bebra affects germ-plasm components:
One of the requirements for germ-cell formation is assembly of the specialized cytoplasm, the germ plasm, at the posterior of the egg. Changing the amount of germ plasm causes a concomitant change in germ-cell number (EPHRUSSI and LEHMANN 1992 ; SMITH et al. 1992 ). Therefore, the integrity of the germ plasm was investigated by analyzing the localization of known germ-plasm components. Table 1 shows the percentage of mutant embryos that show localized osk RNA (from RNA in situ hybridization experiments). The osk mutants used do not show any localization of osk RNA. In l(3)mbt, shkl, and oosy embryos, osk RNA is localized similarly to wild-type embryos, as no decrease in osk localization is seen. Analysis of other germ-plasm components (gcl RNA and pgc RNA) showed similar results (data not shown). In beb mutant embryos, osk RNA localization is reduced, although 31% of embryos still show normal osk localization. This defect in osk RNA localization is also observed during oogenesis in beb mutant females (data not shown). Thus the deficiency in germ-cell formation in beb mutant embryos appears to be a result of germ-plasm defects.

l(3)mbt, shkl, and oosy act independently from or downstream of oskar:
The integrity of the germ plasm in l(3)mbt, shkl, and oosy mutant embryos suggests that the germ cell defect is downstream of or independent of osk function, while beb's defect in osk localization could account for the loss of germ-cell formation. The sufficiency of osk for assembly of germ plasm and formation of germ cells was shown by localizing osk RNA to the anterior of embryos using the bicoid (bcd) 3'-untranslated region (UTR; EPHRUSSI and LEHMANN 1992 ). Since the osk coding region in this transgene is under different RNA localization and translational control, it can also be used to determine whether other genes in the germ-plasm assembly pathway are genetically upstream of or downstream/parallel to osk function. Therefore, the osk-bcd 3'-UTR transgene was crossed into the l(3)mbt, shkl, oosy, and beb backgrounds. Fig 3A shows the phenotype of the osk-bcd 3'-UTR transgene in an otherwise wild-type background. Germ cells form at the posterior due to endogenous osk localization, while germ cells at the anterior are a result of ectopic osk function. If the gene in question functions upstream of the localization of osk RNA, then only the endogenous, posterior germ cells should be affected. If the gene in question functions either downstream of or in parallel with osk localization, then both anterior and posterior germ cells should be affected. Mutations in osk in conjunction with this transgene (see Fig 3B) demonstrate the independence of the anterior germ cells from oskar function (or other upstream components). When the osk-bcd 3'-UTR transgene is crossed into either the l(3)mbt or shkl mutant embryos, germ cells fail to form at both posterior and anterior poles (Fig 3C and Fig D), suggesting that these genes act downstream of or in parallel with osk localization. The oosy mutant embryos with an osk-bcd 3'-UTR transgene show some reduction in germ-cell formation at both poles as well (Fig 3E). beb mutant embryos, on the other hand, still form germ cells at the anterior in the presence of the osk-bcd 3'-UTR transgene while failing to form germ cells posteriorly (Fig 3F). Thus beb appears to act upstream of osk localization.

View larger version (78K): In this window In a new window Download PPT slide   Figure 3. Dependence of maternal germ-cell formation defects on oskar RNA localization. Anterior is to the left. Germ cells are visualized with an anti-Vasa antibody. Arrows point to ectopic anterior germ cells where present. Arrowheads point to endogenous posterior germ cells where present. (A–F) Embryos laid by females containing a osk-bcd 3'UTR (OB) transgene, which causes ectopic osk RNA localization (and thus germ-cell formation) at the anterior. (A) wild type; (B) oskGM52/166, note that the oskGM52 allele is weak and produces some germ cells (arrowhead); (C) l(3)mbtGM76/GM79; (D) shklGM45/GM163; (E) oosyGM47/GM73; (F) bebGM29/GM50. In an otherwise wild-type background, the osk-bcd 3'UTR transgene causes germ cells to form at the anterior (A). Mutations that disrupt germ-cell formation at the posterior without affecting the ectopic anterior germ cells include osk-/- and beb-/- (B and F). Mutations in l(3)mbt, shkl, and oosy cause similar germ-cell defects (within a given embryo) for both the endogenous posterior germ cells and the ectopic anterior germ cells (C–E).

l(3)mbt, shkl, and oosy cause defects in axial expansion, cortical migration, and nuclear division:
Beyond germ-plasm assembly, another critical step in germ-cell formation is the migration of nuclei into the germ plasm just prior to germ-cell formation. Two processes lead to the accumulation of nuclei at the cortex: during nuclear cycles 6–8 nuclei undergo an expansion along the longitudinal axis of the embryo and during cycles 8–9 nuclei migrate to the cortex. The status of the nuclei was investigated in osk, l(3)mbt, shkl, oosy, and beb alleles using the DNA marker Oligreen. osk mutant embryos fail to form germ plasm and thus neither form germ cells nor develop an abdomen. However, the synchronous nuclear divisions are not disrupted and nuclei at cycles 9–10 still migrate to the posterior of the embryos (Fig 4B), with no apparent difference relative to wild type (Fig 4A). In contrast, the synchronous mitotic divisions of the early embryo are disrupted in l(3)mbt embryos (Fig 4C). There are nuclei in almost all stages of mitosis within a single embryo at a given time point, a situation that never occurs in wild type. Gaps in the density of nuclei can sometimes be seen (data not shown), which could result from either migration defects or severe mitotic synchrony disruption. Nuclei in shkl mutant embryos fail to undergo proper axial expansion (Fig 4D and Fig 5D), and as a consequence nuclei fail to reach the posterior of the embryo. oosy mutant embryos have a more extensive nuclear phenotype, in that both axial expansion and nuclear migration to the cortex seem disrupted and the synchrony of the divisions is disturbed (Fig 4E). There is no obvious defect in axial expansion, cortical migration, or synchrony of nuclear divisions in beb mutant embryos (Fig 4F), consistent with a role for beb upstream of osk localization rather than one affecting the nuclei.

View larger version (85K): In this window In a new window Download PPT slide   Figure 4. Early nuclear divisions and migration in embryos laid by mutant females. Anterior is to the left. Nuclei were visualized with OliGreen. (A) wild type; (B) osk-/-; (C) l(3)mbtGM79/Df; (D) shklGM45/GM163; (E) oosyGM47/GM73; (F) bebGM29/GM50. There are no obvious nuclear division or migration defects caused by osk-/- or beb-/- mutations (B and F). Mutations in l(3)mbt disrupt the synchrony of the mitotic divisions (C). Nuclei fail to migrate to the termini, particularly the posterior, in shkl-/- mutants (D). oosy mutations cause mitotic synchrony and migration defects (E).

View larger version (44K): In this window In a new window Download PPT slide   Figure 5. Centrosomes in embryos laid by mutant females. DNA was visualized with OliGreen (green) while centrosomes were visualized with an anti-Centrosomin antibody (red). (A) wild type; (B) l(3)mbtGM79/Df; (C) l(3)mbtE2/Df; (D) shklGM45/GM163; (E) oosyGM47/GM73; (F) bebGM29/GM50. Free centrosomes without associated chromosomes are indicated by arrows, while acentrosomal DNA is indicated by arrowheads. The asynchrony caused by mutations in l(3)mbtGM79 can be seen in B, but no centrosomal defects are evident. Only two nuclei are present in this l(3)mbtE2 mutant embryo where mitotic divisions have stopped (C). Mutations in shkl sometimes cause centrosome loss, but mitotic synchrony appears undisturbed (D). Many centrosomal defects and gaps in nuclear density appear in oosy mutants (E). There are no obvious centrosomal defects in beb mutants (F).

To further understand the nuclear defect, mutant embryos were stained with an antibody against centrosomin, a component of the mitotic centrosome (LI et al. 1998 ). Additionally, an antibody that recognizes a phosphorylated form of histone H3 was used, which recognizes condensed chromosomes only during mitosis (HENDZEL et al. 1997 ). As seen with DNA staining, beb mutant embryos show no defect in centrosomes (Fig 5F) or chromatin condensation (Fig 6E). Nuclei in l(3)mbt mutant embryos seem to have normal mitotic spindles, despite their disrupted mitotic synchrony (Fig 5B). Staining with anti-phospho-Histone H3 starkly reveals the asynchrony of mitotic divisions in l(3)mbt mutant embryos, as domains of nuclei with condensed chromatin are seen (Fig 6B). The pattern and size of these regions are not consistent from embryo to embryo, suggesting a stochastic nature for the asynchrony. However, the fact that phospho-Histone (and therefore condensed chromatin) is present in patches or domains suggests that some mechanism of synchronization is still present in these mutant embryos. Despite the asynchrony, the chromatin in individual nuclei appears normal (Fig 6B, inset). The l(3)mbt^E2 allele shows a stronger phenotype (Fig 5C) in which the nuclei appear to stop dividing and most embryos have only 2–16 nuclei. shkl mutant embryos, beyond the retarded axial migration, show occasional loss of centrosomes (Fig 5D, arrows), and occasional chromosome loss (Fig 6C, inset). The abnormal nuclei seen in oosy mutant embryos are often dissociated from their centrosomes (Fig 5E) and show frequent chromosome loss (Fig 6D). The nuclear perturbations seen in l(3)mbt, shkl, and oosy embryos, coupled with their germ-cell formation defects, suggest that any disruption to the synchronous divisions and migration of the nuclei in the early embryo can have adverse effects on germ-cell formation.

View larger version (92K): In this window In a new window Download PPT slide   Figure 6. Condensed chromatin in embryos laid by mutant females. Anterior is to the left. Chromatin was visualized with an anti-phospho-Histone antibody. (A) wild type; (B) l(3)mbtGM79/Df; (C) shklGM45/GM163; (D) oosyGM47/GM73; (E) bebGM29/GM50. Insets show a close-up of an individual nucleus. The mitotic asynchrony seen in l(3)mbt mutants is evident by the patchiness of phospho-Histone staining. No consistent pattern of the domains of nuclei containing condensed chromatin is present in a collection of l(3)mbt mutant embryos. Individual nuclei appear normal (B). The pattern of nuclei with condensed chromatin in shkl mutants is consistent with a failure of nuclei to migrate to the posterior. Occasional chromosome loss is seen (arrowhead; C). Mutations in oosy cause frequent chromosome loss (arrowhead), asynchronous divisions, and defects in nuclear migration (D). There are no obvious chromosomal defects in beb mutants (E).

l(3)mbt alleles define functional domains and reveal a temperature-sensitive process:
l(3)mbt was originally identified on the basis of a temperature-sensitive phenotype of malignant overgrowth of larval brain cells (GATEFF et al. 1993 ). Sequencing of genomic DNA from the l(3)mbt alleles identified the molecular lesions in both the alleles from this screen as well as those previously identified. Three of the alleles are missense mutations, while the other two are nonsense mutations (see Fig 7). The protein produced from l(3)mbt features several conserved domains (WISMAR et al. 1995 ) and shows homology with proteins in vertebrates (KOGA et al. 1999 ; USUI et al. 2000 ; WISMAR 2001 ; Fig 7). Sex comb on midleg (Scm), a member of the Polycomb group (PcG) of genes, is the closest Drosophila homolog of l(3)mbt (BORNEMANN et al. 1996 ). These proteins share putative Cys₂-Cys₂ zinc fingers whose spacing and sequence conservation define a unique subclass of zinc fingers. They also share malignant brain tumor (MBT) repeats, 74-amino-acid domains with 33–41% identity between the two proteins. The mutations in l(3)mbt^GM79 and l(3)mbt^ts1 alter conserved residues within the MBT domains, demonstrating the functional relevance of the MBT domains. Finally, there is a sterile alpha motif (SAM) domain (also called SPM) with 30% identity between MBT and Scm (Fig 7).

View larger version (9K): In this window In a new window Download PPT slide   Figure 7. Schematics of the Drosophila MBT protein and two related proteins from Drosophila (Scm) and human (KIAA0681). The MBT-repeat domains, putative zinc-fingers, and SAM domains are indicated as shaded blocks. Mutations present in the Drosophila l(3)mbt alleles are indicated by arrowheads (solid arrowheads for nonsense mutations, open arrowheads for missense mutations). Mutations cause amino acid changes as follows: GM76 (W928STOP), GM79 (M964K), E2 (W1099STOP), ts1 (P1130S), and GM161 (H1161Y). Percentage identity between domains is as follows (compared to the corresponding domain from MBT): Scm MBT1, 41%; Scm MBT2, 36%; Scm SAM, 30%; KIAA0681 MBT1, 55%; KIAA0681 MBT2, 45%; KIAA0681 MBT3, 49%; and KIAA0681 SAM, 29%.

In trans to a deficiency that uncovers l(3)mbt [Df(3R)D605 or Df(3R)mbt^PE3], the alleles form a phenotypic series (GM161 < ts1 < GM79 < E2 < GM76) that ranges from a weak version of the maternal-effect mitotic asynchrony described here to a more severe mitotic block (Fig 5C) to oogenesis defects in trans-heterozygous females (data not shown). Additionally, each of these phenotypes is temperature sensitive, in that any allele can be shifted up or down the phenotypic series on the basis of the temperature at which the flies are raised (Fig 8). At 18°, the GM161 and ts1 missense alleles have little or no embryonic defect, while the GM79 missense allele shows slight mitotic asynchrony. The E2 nonsense allele, which causes a mitotic stop at 25° (Fig 5C), completes mitotic cycles at 18°, although they still show significant asynchrony. Females carrying the GM76 nonsense allele at 18° lay embryos with only a few nuclei. These flies do not lay any eggs when raised at 25°. None of the l(3)mbt alleles produce eggs if the flies are raised at 29°, as all have oogenesis defects (data not shown). The larval tumorigenic phenotype originally described for the ts1 allele at the restrictive temperature can also be seen in other allelic combinations (data not shown) with a temperature sensitivity profile similar to that described for l(3)mbt^ts1 (GATEFF et al. 1993 ). The fact that all of the l(3)mbt alleles show temperature sensitivity indicates that mbt mutations somehow render an underlying cellular process temperature sensitive rather than altering each mutant protein to temperature sensitivity.

View larger version (16K): In this window In a new window Download PPT slide   Figure 8. Temperature sensitivity of l(3)mbt alleles. Alleles are listed across the top and phenotypic descriptions are listed at the left. The bars represent the predominant phenotype at a given temperature, tested at 18° (black), 25° (gray), and 29° (white).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Germ-cell formation requires two distinct steps:
The process of germ-cell formation is essential for the reproductive success of the organism. In many organisms, the germ cells are determined early in embryogenesis; in Drosophila, they are the first cells formed. This process is controlled by germ-line-specific determinants, accumulating in a specialized cytoplasm called germ plasm. In Drosophila, the germ plasm is established during oogenesis, in a process that is dependent upon oskar. Failure to form an intact germ plasm, through mutations in the osk locus or by other means, results in a failure to form germ cells in the developing embryo (WILLIAMSON and LEHMANN 1996 ).

A second process is also required for normal germ-cell formation in Drosophila. The early embryo is a nuclear syncytium, a single large cell that undergoes 13 rounds of synchronous mitoses without cell division. The migration of these nuclei to the periphery of the embryo begins at the eighth mitotic cycle, when some nuclei move toward the posterior and the germ plasm. The nuclei and germ plasm cellularize at the posterior, forming PGCs. Mutants defective in this migration of nuclei to the posterior of the embryo fail to form germ cells (NIKI 1984 ).

These two processes, germ-plasm formation and nuclear migration, while both required for germ-cell formation, are parallel and independent pathways. In mutants defective in axial expansion, nuclei fail to reach the posterior at the appropriate time [shkl, reported here and gs(1)N26 and gs(1)N441 (NIKI and OKADA 1981 )], the germ plasm is intact, but the germ-plasm components do not remain stable (IIDA and KOBAYASHI 2000 ). This breakdown occurs after these components would normally be sequestered from cytoplasmic degradation pathways because of germ-cell cellularization. Germ-plasm integrity is not dependent upon the migration of nuclei until after the time when cellularization would have normally occurred. If cellularization fails and the germ plasm is not sequestered in the germ cells, the localized components of the germ plasm either are degraded or become delocalized (IIDA and KOBAYASHI 2000 ). Additionally, the migration of nuclei to the posterior is not dependent upon proper germ-plasm formation. In osk mutant embryos, where germ plasm does not form (LEHMANN and NUSSLEIN-VOLHARD 1986 ), germ cells do not form. This is not due, however, to any defect in nuclear migration. At mitotic cycle nine, nuclei have moved to the posterior of osk embryos in a manner that is indistinguishable from wild-type embryos. beb mutant embryos are similar in that the nuclei divide and migrate normally although the germ plasm is compromised. Thus the germ plasm plays no role in axial expansion or cortical migration of the syncytial nuclei and the two processes required for germ-cell formation, namely germ-plasm formation and nuclear migration, act independently in the early embryo.

Germ-plasm components downstream of oskar:
Mutations that affect the formation of the germ cells have generally been identified on the basis of a phenotype distinct from germ-cell formation, one affecting the second function of germ plasm: abdominal patterning. The only exception are the two grandchildless(gs)-like genes (NIKI and OKADA 1981 ), but those two mutants affect axial expansion and nuclear migration (as in shkl), not germ plasm. Mutations that affect formation of the germ plasm generally also have defects in abdominal patterning as both processes are dependent upon localization of other germ-plasm components through osk function. Failure to localize nos RNA in an intact germ plasm causes embryonic patterning defects (WANG and LEHMANN 1991 ; GAVIS and LEHMANN 1992 ). Those germ-plasm components that are downstream of oskar and specific for germ-cell formation were identified through molecular or other nongenetic means. Mitochondrial large rRNA was identified through rescue (by injection of the RNA) of the germ-cell formation defect caused by UV irradiation (KOBAYASHI and OKADA 1989 ). The germ-cell-less (gcl) gene was identified serendipitously via its RNA localization pattern (JONGENS et al. 1992 ). The screen described here was designed to identify components of the germ plasm that are specific for germ-cell formation. While new alleles of some genes known to be involved in this process were identified, and beb mutants appear to affect germ-plasm localization and therefore germ-cell formation, no new germ-cell-specific genes downstream of osk function were identified. The screen covered 20% of the genome (chromosome 3R). This does suggest that the number of germ-cell-specific genes that can be identified using our screening protocol has to be rather small. The pathway leading to PGC formation downstream of osk function currently includes the large ribosomal RNA from mitochondria (IIDA and KOBAYASHI 1998 ) and a novel protein that localizes to the nucleoplasmic surface of the nuclear envelope (GCL; JONGENS et al. 1994 ). It seems likely that additional components will be required, so it is somewhat surprising that none were uncovered here. Perhaps many of the RNAs and proteins required for germ-cell formation have additional roles in oogenesis or cellularization that would prohibit their discovery in this type of screen.

Perturbations in cell cycle or nuclear migration specifically affect germ-cell formation:
The three genes identified in this screen [l(3)mbt, shkl, and oosy] affect the axial and cortical nuclear migration or the synchrony of the nuclear division cycles in the early embryo. Our mutant analysis shows that defects in these processes lead to a specific phenotype in germ-cell formation because of the precise temporal and spatial requirements placed on the nuclei in this process. The most prominent defect we observe in embryos from females mutant for l(3)mbt is a disruption in the synchronous mitotic cycles of the early embryo. The strongest allelic combinations of l(3)mbt alleles show defects in the divisions already at the onset of embryonic development. It is thus likely that nuclear division defects are the primary cause of the subsequent defects in synchrony of nuclear division, nuclear migration, and germ-cell formation in l(3)mbt mutants. The primary defect in shkl mutants seems to lie in a failure of nuclei to reach the posterior pole during axial expansion. This process has been shown to be actin dependent (HATANAKA and OKADA 1991 ). Interestingly, mutations in spaghetti squash (sqh), the Drosophila myosin II, also exhibit defects in axial expansions and asymmetry of nuclear division, similar to those observed in shkl mutants (WHEATLEY et al. 1995 ). oosy mutant embryos exhibit defects in axial and cortical nuclear migration but also show severe defects in nuclear divisions with centrosome defects and chromosome loss. This suggests that in this mutant, similar to l(3)mbt, nuclear migration defects may be the consequence rather than the cause of defects in nuclear division.

l(3)mbt, oosy, and shkl mutant embryos generally recover from their somatic defects. Somatic cellularization still occurs, and most embryos go on to develop and hatch. When mitotic arrest or nuclear loss is induced in embryos by UV irradiation or DNA injection, a majority of the embryos still cellularize and form a cuticle (YASUDA et al. 1991 ). Despite the unusual plasticity of the early embryo and its ability to recover from dramatic nuclear defects, the ability to form germ cells is restricted to a specific time and place. The sensitivity of germ-cell formation to disruptions in nuclear migration and the early embryonic cell cycle may thus provide an efficient and reliable tool to uncover genes regulating the actin and microtubule network or the cell cycle in the early Drosophila embryo.

Despite the different types of nuclear defects caused by mutations described here, the numbers of germ cells that do form in each mutant line are similar (Table 1). Additionally, these germ cells are fully functional, as most embryos develop into fertile adults. While l(3)mbt and shkl mutants show little or no chromosomal or centrosomal aberrations (Fig 5B and Fig D, and Fig 6B and Fig C), nuclei in oosy embryos frequently lose chromosomes or centrosomes (Fig 5E and Fig 6D), but are still able to form germ cells. This is not too surprising since centrosomes alone have been proposed to be able to induce germ-cell formation (RAFF and GLOVER 1989 ). However, the fact that the germ cells formed in oosy embryos are functional in the adult may suggest that the most critical factor in germ-cell formation is the timing and presence of nuclei and centrosomes and that some defects in mitosis may be tolerated.

The cyclins are known to regulate the early mitotic divisions, and perturbations in the levels of CyclinB have been shown to also affect the speed of nuclear migration via its effects on microtubule dynamics (STIFFLER et al. 1999 ; JI et al. 2002 ). Indeed, the phenotypes of embryos from mothers with six copies of a CyclinB transgene show nuclear asynchrony defects similar to those observed in l(3)mbt, shkl, and oosy mutant embryos (JI et al. 2002 ). Preliminary experiments in which the level of Cyclin A, B, or B3 was changed (via hemizygous Cyclin mothers or maternal addition of transgenic genomic copies) in a l(3)mbt mutant showed no effect on the maternally determined asynchrony defect (data not shown). Additional investigations into the effects of these mutants on Cyclin levels and regulation, as well as potential interactions with Cyclins or the Cyclin regulatory machinery, may elucidate the molecular nature of the perturbations in the synchronous divisions and migration in the early embryo.

l(3)mbt as a model for tumorigenesis:
l(3)mbt was originally identified as a tumor suppressor (GATEFF 1978 ). The temperature-sensitive allele ts1 caused malignant overgrowth of the larval brain and imaginal disc overproliferation at the restrictive temperature (GATEFF et al. 1993 ). Interestingly, all l(3)mbt alleles are temperature sensitive, including the nonsense alleles, which would produce a truncated and potentially unstable protein. This suggests that MBT protein regulates a temperature-sensitive process, rather than temperature sensitivity being specific to any allele. Progression of the cell cycle in yeast is known to be temperature sensitive at several stages, and heat-shock proteins have been shown to regulate Cyclin complexes (ZHU et al. 1997 ). Thus, MBT may target or protect one of these temperature-sensitive stages of the cell cycle.

Scm is the closest Drosophila homolog of MBT (Fig 7; BORNEMANN et al. 1996 ). Scm is a member of the PcG of genes. Embryos lacking both maternal and zygotic Scm product undergo a homeotic transformation of most segments to the eighth abdominal segment (BREEN and DUNCAN 1986 ). l(3)mbt does not appear to be a PcG gene, as mutant alleles do not cause homeotic transformations. Analysis of the protein domains conserved between Scm and MBT protein may help to reveal the functional properties of MBT. The SAM domain of Scm has been shown to mediate heterotypic and homotypic protein-protein interactions with the polyhomeotic (ph) gene product, another PcG gene (PETERSON et al. 1997 ). Analysis of Scm mutant alleles with alterations in the MBT repeats suggests that, while these protein domains are not crucial for the PcG protein-protein interaction or targeting of the protein, they are crucial for the protein's biochemical function (BORNEMANN et al. 1998 ). While the function of MBT domains remains elusive, MBT repeats are found in vertebrate proteins. KIAA0681 is a human protein that contains three MBT repeats (45–55% identical to those in the Drosophila MBT protein), a putative Cys₂-Cys₂ zinc finger, and an SAM domain (29% identical to the SAM domain in Drosophila MBT; Fig 7). This protein associates with chromatin only during mitosis in human cultured cell lines, appearing scattered in interphase nuclei. Upon overexpression of the human l(3)mbt homolog in a human glioma cell line, nuclear segregation and cytokinesis are affected and multinucleated cells are formed (KOGA et al. 1999 ). This protein could be a l(3)mbt ortholog, functioning in vertebrate cell-cycle control as it does in Drosophila. Given l(3)mbt's role as a tumor suppressor in a model organism with powerful genetics, further study may lead to an understanding of the cell cycle's role in tumorigenesis that can be applied to human biology.

	FOOTNOTES

¹ Present address: Bioinformatics Scientist, Favrille, Inc., San Diego, CA 92121.

	ACKNOWLEDGMENTS

We thank Thomas Kaufman for Centrosomin antibody and Steven Wasserman for sharing detailed molecular information from the l(3)mbt genomic region. We are grateful for fly stocks from Jasmine Wismar, including deficiencies near l(3)mbt and the previously published l(3)mbt alleles. We are grateful to the Bloomington Drosophila stock collection for sending us many mutant lines and the deficiency kit. In addition, we thank members of the Lehmann laboratory for intellectual contributions to this manuscript. Finally, we are grateful to the members of the 3R screen team. C.B.Y. was the recipient of National Institutes of Health postdoctoral fellowship HD08263-03 and was a Howard Hughes Medical Institute (HHMI) Associate. R.L. is an HHMI investigator.

Manuscript received January 7, 2003; Accepted for publication August 13, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ASHBURNER, M., 1989a Mutation and mutagenesis, p. 380 in Drosophila: A Laboratory Handbook, edited by M. ASHBURNER. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ASHBURNER, M., 1989b Preparation of DNA from single flies, pp. 108–109 in Drosophila: A Laboratory Manual, edited by M. ASHBURNER. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BAKER, J., W. E. THEURKAUF, and G. SCHUBIGER, 1993  Dynamic changes in microtubule configuration correlate with nuclear migration in the preblastoderm Drosophila embryo. J. Cell Biol. 122(1):113-121.[Abstract/Free Full Text]

BORNEMANN, D., E. MILLER, and J. SIMON, 1996  The Drosophila Polycomb group gene Sex comb on midleg (Scm) encodes a zinc finger protein with similarity to polyhomeotic protein. Development 122:1621-1630.[Abstract]

BORNEMANN, D., E. MILLER, and J. SIMON, 1998  Expression and properties of wild-type and mutant forms of the Drosophila sex comb on midleg (SCM) repressor protein. Genetics 150:675-686.[Abstract/Free Full Text]

BREEN, T. R. and I. M. DUNCAN, 1986  Maternal expression of genes that regulate the bithorax complex of Drosophila melanogaster. Dev. Biol. 118:442-456.[Medline]

CHOU, T. B. and N. PERRIMON, 1996  The autosomal FLP-DFS technique for generating germline mosaics in Drosophila melanogaster. Genetics 144:1673-1679.[Abstract]

EDGAR, B. A., F. SPRENGER, R. J. DURONIO, P. LEOPOLD, and P. H. O'FARRELL, 1994  Distinct molecular mechanism regulate cell cycle timing at successive stages of Drosophila embryogenesis. Genes Dev. 8:440-452.[Abstract/Free Full Text]

EPHRUSSI, A. and R. LEHMANN, 1992  Induction of germ cell formation by oskar. Nature 358:387-392.[Medline]

EPHRUSSI, A., L. K. DICKINSON, and R. LEHMANN, 1991  Oskar organizes the germ plasm and directs localization of the posterior determinant nanos. Cell 66:37-50.[Medline]

ERDELYI, M., A. M. MICHON, A. GUICHET, J. B. GLOTZER, and A. EPHRUSSI, 1995  Requirement for Drosophila cytoplasmic tropomyosin in oskar mRNA localization. Nature 377(6549):524-527.[Medline]

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

FOE, V. E and B. M. ALBERTS, 1983  Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis. J. Cell Sci. 61:31-70.[Abstract]

GATEFF, E., 1978  Malignant neoplasms of genetic origin in Drosophila melanogaster. Science 200:1448-1459.[Abstract/Free Full Text]

GATEFF, E., T. LOFFLER, and J. WISMAR, 1993  A temperature-sensitive brain tumor suppressor mutation of Drosophila melanogaster: developmental studies and molecular localization of the gene. Mech. Dev. 41:15-31.[Medline]

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

GOLIC, K. G., 1991  Site-specific recombination between homologous chromosomes in Drosophila. Science 252:958-961.[Abstract/Free Full Text]

GOLIC, K. G. and S. LINDQUIST, 1989  The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 59:499-509.[Medline]

GRETHER, M. E., J. M. ABRAMS, J. AGAPITE, K. WHITE, and H. STELLER, 1995  The head involution defective gene of Drosophila melanogaster functions in programmed cell death. Genes Dev. 9:1694-1708.[Abstract/Free Full Text]

HATANAKA, K. and M. OKADA, 1991  Retarded nuclear migration in Drosophila embryos with aberrant F-actin reorganization caused by maternal mutations and by cytochalasin treatment. Development 111:909-920.[Abstract/Free Full Text]

HENDZEL, M. J., Y. WEI, M. A. MANCINI, A. VAN HOOSER, and T. RANALLI et al., 1997  Mitosis-specific phosphorylation of histone H3 initiates primarily within pericentromeric heterochromatin during G2 and spreads in an ordered fashion coincident with mitotic chromosome condensation. Chromosoma 106:348-360.[Medline]

IIDA, T. and S. KOBAYASHI, 1998  Essential role of mitochondrially encoded large rRNA for germ-line formation in Drosophila embryos. Proc. Natl. Acad. Sci. USA 95:11274-11278.[Abstract/Free Full Text]

IIDA, T. and S. KOBAYASHI, 2000  Delocalization of polar plasm components caused by grandchildless mutations, gs(1)N26 and gs(1)N441, in Drosophila melanogaster.. Dev. Growth Differ. 42:53-60.[Medline]

JACOBS, H. W., J. A. KNOBLICH, and C. F. LEHNER, 1998  Drosophila Cyclin B3 is required for female fertility and is dispensable for mitosis like Cyclin B. Genes Dev. 12:3741-3751.[Abstract/Free Full Text]

JI, J. Y., M. HAGHNIA, C. TRUSTY, L. S. GOLDSTEIN, and G. SCHUBIGER, 2002  A genetic screen for suppressors and enhancers of the Drosophila cdk1-cyclin B identifies maternal factors that regulate microtubule and microfilament stability. Genetics 162:1179-1195.[Abstract/Free Full Text]

JONGENS, T. A., B. HAY, L. Y. JAN, and Y. N. JAN, 1992  The germ cell-less gene product: a posteriorly localized component necessary for germ cell development in Drosophila. Cell 70:569-584.[Medline]

JONGENS, T. A., L. D. ACKERMAN, J. R. SWEDLOW, L. Y. JAN, and Y. N. JAN, 1994  Germ cell-less encodes a cell type-specific nuclear pore-associated protein and functions early in the germ-cell specification pathway of Drosophila. Genes Dev. 8:2123-2136.[Abstract/Free Full Text]

KIM-HA, J., J. L. SMITH, and P. M. MACDONALD, 1991  oskar mRNA is localized to the posterior pole of the Drosophila oocyte. Cell 66:23-35.[Medline]

KOBAYASHI, S. and M. OKADA, 1989  Restoration of pole-cell-forming ability to u.v.-irradiated Drosophila embryos by injection of mitochondrial lrRNA. Development 107:733-742.[Abstract/Free Full Text]

KOGA, H., S. MATSUI, T. HIROTA, S. TAKEBAYASHI, and K. OKUMURA et al., 1999  A human homolog of Drosophila lethal(3)malignant brain tumor (l(3)mbt) protein associates with condensed mitotic chromosomes. Oncogene 18:3799-3809.[Medline]

LEHMANN, R. and C. NUSSLEIN-VOLHARD, 1986  Abdominal segmentation, pole cell formation, and embryonic polarity require the localized activity of oskar, a maternal gene in Drosophila. Cell 47:141-152.[Medline]

LEHMANN, R. and D. TAUTZ, 1994  In situ hybridization to RNA. Methods Cell Biol. 44:575-598.[Medline]

LI, K., E. Y. XU, J. K. CECIL, F. R. TURNER, and T. L. MEGRAW et al., 1998  Drosophila centrosomin protein is required for male meiosis and assembly of the flagellar axoneme. J. Cell Biol. 141:455-467.[Abstract/Free Full Text]

LÖFFLER, T. W., 1988 Genetic, cytogenetic and molecular genetic investigation of the recessive tumor suppressor gene lethal(3)malignant brain tumor of Drosophila melanogaster. Ph.D. Thesis, Johannes Gutenberg University, Mainz, Germany.

MOORE, L. A., H. T. BROIHIER, M. VAN DOREN, L. B. LUNSFORD, and R. LEHMANN, 1998  Identification of genes controlling germ cell migration and embryonic gonad formation in Drosophila. Development 125:667-678.[Abstract]

MORRIS, J. Z., C. NAVARRO and R. LEHMANN, 2003 Identification and analysis of mutations in bob, Doa and eight new genes required for oocyte specification and development in Drosophila melanogaster. Genetics 164: 1435–1446.

NIKI, Y., 1984  Developmental analysis of the grandchildless (gs(1)N26) mutation in Drosophila melanogaster: abnormal cleavage patterns and defects in pole cell formation. Dev. Biol. 103:182-189.[Medline]

NIKI, Y. and M. OKADA, 1981  Isolation and characterization of grandchildless-like mutants in Drosophila melanogaster.. Wilhelm Roux's Arch. 190:1-10.

NUSSLEIN-VOLHARD, C., H. G. FROHNHOFER, and R. LEHMANN, 1987  Determination of anteroposterior polarity in Drosophila. Science 238:1675-1681.[Abstract/Free Full Text]

PETERSON, A. J., M. KYBA, D. BORNEMANN, K. MORGAN, and H. W. BROCK et al., 1997  A domain shared by the Polycomb group proteins Scm and ph mediates heterotypic and homotypic interactions. Mol. Cell. Biol. 17:6683-6692.[Abstract]

PRITCHARD, D. K. and G. SCHUBIGER, 1996  Activation of transcription in Drosophila embryos is a gradual process mediated by the nucleocytoplasmic ratio. Genes Dev. 10:1131-1142.[Abstract/Free Full Text]

RAFF, J. W. and D. M. GLOVER, 1989  Centrosomes, and not nuclei, initiate pole cell formation in Drosophila embryos. Cell 57:611-619.[Medline]

ROYOU, A., W. SULLIVAN, and R. KARESS, 2002  Cortical recruitment of nonmuscle myosin II in early syncytial embryos: its role in nuclear axial expansions and its regulation by cdc2 activity. J. Cell Biol. 158:127-137.[Abstract/Free Full Text]

SEYDOUX, G. and M. A. DUNN, 1997  Transcriptionally repressed germ cells lack a subpopulation of phosphorylated RNA polymerase II in early embryos of Caenorhabditis elegans and Drosophila melanogaster. Development 124(11):2191-2201.[Abstract]

SMITH, J. L., J. E. WILSON, and P. M. MACDONALD, 1992  Overexpression of oskar directs ectopic activation of nanos and presumptive pole cell formation in Drosophila embryos. Cell 70(5):849-859.[Medline]

STIFFLER, L. A., J. Y. JI, S. TRAUTMANN, C. TRUSTY, and G. SCHUBIGER, 1999  Cyclin A and B functions in the early Drosophila embryo. Development 126:5505-5513.[Abstract]

ST. JOHNSTON, D. and C. NUSSLEIN-VOLHARD, 1992  The origin of pattern and polarity in the Drosophila embryo. Cell 68(2):201-219.[Medline]

USUI, H., T. ICHIKAWA, K. KOBAYASHI, and T. KUMANISHI, 2000  Cloning of a novel murine gene Sfmbt, Scm-related gene containing four mbt domains, structurally belonging to the Polycomb group of genes. Gene 248:127-135.[Medline]

VAN DOREN, M., A. L. WILLIAMSON, and R. LEHMANN, 1998  Regulation of zygotic gene expression in Drosophila primordial germ cells. Curr. Biol. 8:243-246.[Medline]

VAN EEDEN, F. J., I. M. PALACIOS, M. PETRONCZKI, M. J. WESTON, and D. ST. JOHNSTON, 2001  Barentsz is essential for the posterior localization of oskar mRNA and colocalizes with it to the posterior pole. J. Cell Biol. 154(3):511-523.[Abstract/Free Full Text]

WANG, C. and R. LEHMANN, 1991  Nanos is the localized posterior determinant in Drosophila. Cell 66:637-647.[Medline]

WHEATLEY, S., S. KULKARNI, and R. KARESS, 1995  Drosophila nonmuscle myosin II is required for rapid cytoplasmic transport during oogenesis and for axial migration in the early embryos. Development 121:1937-1946.[Abstract]

WILLIAMSON, A. and R. LEHMANN, 1996  Germ cell development in Drosophila. Annu. Rev. Cell Dev. Biol. 12:365-391.[Medline]

WISMAR, J., 2001  Molecular characterization of h-l(3)mbt-like: a new member of the human mbt family. FEBS Lett. 507:119-121.[Medline]

WISMAR, J., T. LOFFLER, N. HABTEMICHAEL, O. VEF, and M. GEISSEN et al., 1995  The Drosophila melanogaster tumor suppressor gene lethal(3)malignant brain tumor encodes a proline-rich protein with a novel zinc finger. Mech. Dev. 53:141-154.[Medline]

YASUDA, G. K., J. BAKER, and G. SCHUBIGER, 1991  Temporal regulation of gene expression in the blastoderm Drosophila embryo. Genes Dev. 5:1800-1812.[Abstract/Free Full Text]

ZALOKAR, M. and I. ERK, 1976  Division and migration of nuclei during early embryogenesis of Drosophila melanogaster. J. Microsc. Biol. Cell. 25:97-106.

ZHU, D., D. J. DIX, and E. M. EDDY, 1997  HSP70–2 is required for CDC2 kinase activity in meiosis I of mouse spermat