articles

Activation of a meiotic checkpoint regulates translation of Gurken during Drosophila oogenesis Amin Ghabrial* and Trudi Schüpbach*† *HHMI, Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544, USA †e-mail: [email protected]

The genes okra and spindle-B act during meiosis in Drosophila to repair double-stranded DNA breaks (DSBs) associated with meiotic recombination. Unexpectedly, mutations in these genes cause dorsoventral patterning defects during oogenesis. These defects result from a failure to accumulate Gurken protein, which is required to initiate dorsoventral patterning during oogenesis. Here we find that the block in Gurken accumulation in the oocyte cytoplasm reflects activation of a meiotic checkpoint in response to the persistence of DSBs in the nucleus. We also show that Vasa is a target of this meiotic checkpoint, and so may mediate the checkpoint-dependent translational regulation of Gurken.

354

Results

Genomic competitor:

1x

spn-B

Ore-R

DNA ladder

Given Unrepaired DSBs cause dorsoventral patterning defects. that yeast genes homologous to okr and spn-B are known to function in the nucleus, where their encoded proteins bind to DNA and catalyse DNA-strand exchange as part of the recombination-repair process8, it is surprising that, in Drosophila, mutations in these genes cause defects in Grk accumulation in the oocyte cytoplasm. We first tested whether the production of eggs with dorsoventral patterning defects by mutant females reflects a defect in repair of DSBs, or reveals a new function for these genes. In Saccharomyces cerevisiae, mutations that prevent DSB repair during meiotic recombination can be suppressed by mutations in SPO11 (reviewed in ref. 9), which encodes a topoisomerase-II-like protein that is required to make the DSBs that initiate meiotic recombination10,11.

3x

1x

3x

bp 529 392

Genomic competitor:

1x

3x

1x

okr

spn-C

Ore-R

DNA ladder

M

utations in okra (okr), spindle-B (spn-B) and spindle-C (spnC) are members of a group of female-sterile mutations (spindle class) that produce variable eggshell defects in Drosophila. Egg chambers mutant for okr, spn-B or spn-C fail to accumulate wild-type levels of Gurken (Grk) — a transforming growth factor-α-like signalling molecule that is used to communicate information on oocyte polarity to adjacent follicle cells1,2. In situ hybridization analysis has shown that grk messenger RNA is present in these mutant oocytes1,2, and, using competitive polymerase chain reaction (PCR), we have confirmed that grk mRNA levels are unaffected (Fig. 1). Reduced levels of Grk protein lead to the production of eggshells and embryos with anterior/posterior and dorsal/ventral patterning defects (reviewed in ref. 3). In addition to these patterning defects, mutations in genes of the spindle class also cause defects in the appearance of the oocyte nucleus. DNA within the germinal vesicle of wild-type egg chambers assumes a highly compacted spherical morphology by stage 3 of oogenesis, whereas the DNA within spindle-class-mutant oocyte nuclei is often fragmented or thread-like in appearance1,2. As mutations in vasa, which plays no part in DSB repair, produce a similar nuclear-morphology phenotype4,5, it is unlikely that the fragmented appearance of the DNA in vasa and spindle-class mutants directly reflects unrepaired DSBs. It has been suggested that the appearance of the DNA in these mutants resembles the appearance of DNA in earlier stages of oogenesis2, raising the possibility that the karyosome defect is indicative of delayed progression through meiosis. spn-B, spn-C and okr all have roles in DNA repair. spn-B (DMC1/RAD51-like) and okr (Dmrad54) are homologous to genes in the yeast RAD52 epistasis group that function in the recombinational repair of DSBs, and in Drosophila mutations in these genes lead to mitotic and/or meiotic defects, consistent with a requirement in DNA repair1,6. Because of the phenotypic similarities between spn-C and okr and spn-B mutants, it seemed likely that spnC might have a similar function. Indeed, through complementation tests, we have found that spn-C is allelic to mutagen-sensitive 301 (mus301) (Tables 1, 2 and Methods), which confers sensitivity to methyl methane sulphonate and causes an increase in the frequency of X-chromosome non-disjunction7. Here we analyse the effects of mutations in spn-B, spn-C and okr in combination with mutations in genes required for initiation of meiotic recombination or for a checkpoint that delays progression through meiosis until DSBs are repaired. We find that spn-B, spn-C or okr mutations result in the activation of a meiotic DNA-repair checkpoint, and that this checkpoint may act through Vasa to regulate Grk protein levels.

3x

1x

3x

bp 529 392

Figure 1 Competitive PCR analysis of grk mRNA levels. Single ovaries dissected from freshly eclosed females were used in reverse-transcription reactions as described5. Ore-R, wild-type females; spn-B, spn-C and okr, females with mutations in spn-B, spn-C and okr, respectively. Normalized samples were tested for relative amounts of grk mRNA. Genomic competitor DNA was added in threefold increments until the primary PCR product shifted from the 392-base-pair (bp) band amplified from cDNA derived from reverse-transcribed grk mRNA to the 529-bp band amplified from the genomic competitor. All samples underwent this transition over the same range of genomic competitor, indicating that grk mRNA levels were roughly the same in these ovaries.

NATURE CELL BIOLOGY | VOL 1 | OCTOBER 1999 | cellbio.nature.com

articles a

b

Table 1 Suppression of mutant eggshell phenotypes by mei-W68 Maternal genotype (n)

Wildtype

mei-W68 or Df/CyO; 2% spn-B153/spn-BBU (417)

c

e

f

100 Grk-positive (%)

i

80 60 40 20 0

26%

66%

0%

99%

<1%

<1%

0%

0%

mei-W68 or Df/CyO; 17% spn-C094/mus301D4 (453)

15%

38%

30%

0%

mei-W68/Df; spn-C094/mus301D4 (689)

99.9% 0%

<1%

0%

0%

okrAA/okrRU (1,030)

71%

15%

12%

1%

31%

okrAA mei-W68/ okrRU Df (1,054)

99.6% <1%

<1%

0%

0%

The dorsal side of a Drosophila egg is characterized by the presence of two dorsal appendages, one on each side of the dorsal midline. Classification of eggshells on the basis of dorsoventral polarity was done according to the following criteria: wild-type, two distinct dorsal appendages of wild-type morphology; weakly ventralized, loss of dorsal-most eggshell structures resulting in a partial fusion of the dorsal appendages; moderately ventralized, complete fusion of the dorsal appendages; strongly ventralized, loss of all dorsal structures resulting in the absence of dorsal appendage material. Anteroposterior polarity of eggshells was evaluated by inspection of the posterior ends of eggs for the duplication of the anterior structure, called a micropyle. CyO is a balancer chromosome and carries a wild-type allele at the mei-W68 locus. Df, Df(2R)LL5, which contains a deletion that removes mei-W68 (ref. 12).

h

Normal karyosome (%)

g

6%

mei-W68/Df; spn-B153/spn-BBU (749)

d

AnteroWeakly Moderately Strongly posterior ventralized ventralized ventralized defect

spn-B153

spn-B BU

mei-W68; spn-B153

mei-41; spn-B BU

Figure 2 Confocal analysis of karyosome morphology and Grk accumulation in wild-type and mutant Drosophila egg chambers. Dorsal side is up, and anterior is to the left. The egg chambers consist of a cluster of anterior nurse cells (to the left) and one posteriorly situated oocyte (to the right) surrounded by a layer of follicle cells. In a, c, e, g, DNA (green) and cortical actin (red) are shown. Insets show an enlarged view of the oocyte nucleus. In b, d, f, h, Grk protein (green) and cortical actin (red) are shown (overlap in yellow). a, b, Wild-type egg chambers (OreR). Note the disk-like shape of the DNA in the confocal section in a, and the presence of Grk protein at the dorsal anterior cortex of the oocyte in b. c, d, spn-B mutant egg chambers. Note the fragmented appearance of the karyosome in c and the absence of Grk staining in d. e, f, mei-W68 spn-B double-mutant egg chambers. Note the restoration of normal karyosome morphology and Grk staining. g, h, mei-41 spn-B double-mutant egg chambers. The karyosome and the Grk staining appear normal. i, Confocal sections were studied to determine the frequency with which normally compacted spherical karyosomes and Grk protein could be detected. At least 20 egg chambers were scored for each genotype.

In Drosophila, mei-W68 has been identified as the SPO11 homologue12. Accordingly, we made flies doubly mutant for meiW68 and either okr, spn-B or spn-C to determine whether spindleclass dorsoventral eggshell patterning defects would be produced in the absence of DSB formation. In double-mutant egg chambers, Grk protein accumulates in the same way as in wild-type egg chambers (Fig. 2b, d, f, i). Moreover, patterning of the eggshells, which is NATURE CELL BIOLOGY | VOL 1 | OCTOBER 1999 | cellbio.nature.com

very sensitive to the levels of Grk protein, proceeds normally in double-mutant females (Table 1). Other spindle-class defects are also suppressed by mei-W68: oocyte nuclear morphology in double-mutant egg chambers appears normal, and fertility of doublemutant females is comparable to that observed for females mutant only for mei-W68 (Fig. 2a, c, e, i and data not shown). Similarly, mutations in mei-P22 and c(3)G that, like mei-W68 mutations, eliminate meiotic recombination13,14 are able to suppress okr mutant phenotypes (data not shown). Thus, in the absence of DSBs, the developmental defects observed in okr, spn-B and spn-C mutants are suppressed, indicating that these phenotypes are indeed caused by the presence of improperly processed DSBs during meiosis. The defect in Grk accumulation is checkpoint dependent. Next, we sought to determine the mechanism through which the persistence of unrepaired DSBs during oogenesis causes patterning defects. In S. cerevisiae, mutations in the DSB-repair genes RAD51 and DMC1 cause cells to arrest in prophase of the first meiotic division15. Meiotic arrest is also observed in Dmc1-deficient mice16. In S. cerevisiae, this meiotic arrest is checkpoint dependent — a failure to repair DSBs does not cause meiotic arrest directly; instead, recognition of the presence of DSBs triggers the activation of a pathway that halts the cell cycle until the damaged DNA is repaired. One of the checkpoint genes required to arrest meiosis in response to the presence of unrepaired DSBs is MEC1, which encodes a member of the ATM/ATR subfamily of phosphatidylinositol-3OH-kinase-like proteins17. In otherwise wild-type yeast, mec1 mutations lead to occasional premature meiosis I, as indicated by the persistence of foci of Rad51, which are believed to represent sites of DSB repair, on metaphase-I chromosomes17. These results indicate that Mec1 may normally act to delay the cell cycle in the presence of repair intermediates. In Drosophila, mei-41 encodes a homologue of Mec1, and mei-41 mutants show meiotic non-disjunction as well as maternal-effect defects in the timing of mitotic cell cycles in the early embryo (reviewed in refs 18, 19). To test whether the production of patterning defects by mutations in the spindle-class genes is due to the engagement of an analogous meiotic checkpoint, we have made flies doubly mutant for okr, spn-B or spn-C and mei-41. Mutations in mei-41 are indeed able 355

articles Nucleus Initiation of meiotic recombination Region 2 & 3 of germarium (mei-W68)

Cytoplasm

Stage of oogenesis

spn-B

spn-B

Wild-type

a

mei-41; spn-B

spn-B

vas

Wild-type

b

spn-B

mei-W68; spn-B

Checkpoint (mei-41) DSBs

Translational regulator ? (vas)

Recombination repair (okr, spn-B, spn-C, spn-D)

Double Holliday junctions Translational targets

Figure 3 Western blots of Vasa protein in wild-type and mutant ovarian extracts. a, Migration of Vas isolated from wild-type (Ore-R) ovarian extracts (arrow) and from spn-B mutant ovarian extracts (arrowhead). Vas migration is retarded in spn-B extracts. b, Vas from mei-W68 spn-B and mei-41 spn-B extracts migrates at the lower relative molecular mass (Mr value) seen in wild-type extracts, rather than at the higher Mr value observed in spn-B extracts.

Karyosome – stage 3 A/P axis – stage 3-6

Maternal genotype (n)

Wildtype

Weakly ventralized

Moderately ventralized

Strongly ventralized

spn-BBU/spn-BBU (316)

21%

2%

3%

74%

mei-41 or /ClB; spn-BBU/spn-BBU (1,005)

40%

25%

23%

13%

mei-41RT1/mei-41D3; spn-BBU/spn-BBU (1,000)

95%

3%

1%

1%

spn-C094/mus301D4 (603)

13%

25%

31%

30%

mei-41 /ClB; spn-C094/mus301D4 (576)

30%

23%

27%

20%

mei-41D3/mei-41D3; spn-C094/mus301D4 (978)

97%

2%

1%

<1%

okrRU/okrRU (866)

45%

22%

31%

1%

mei-41 /mei-41 ; okrRU/okrRU (872)

95%

2%

2%

1%

mei-41D3/ClB; vasRG/vasPH165 (1,000)

2%

3%

36%

59%

mei-41D3/mei-41D3; vasRG/vasPH (575)

3%

2%

35%

60%

D3

RT1

D3

D1

D11

Classification of dorsoventral patterning defects in eggshells was as described in Table 1. ClB is a balancer chromosome and carries a wild-type allele at the mei-41 locus.

to suppress the dorsoventral patterning defects caused by mutations in the spindle-class genes. In double-mutant flies, we observed a dramatic increase in the accumulation of Grk protein, as indicated by whole-mount antibody staining (Fig. 2h, i) and by restoration of anteroposterior and dorsoventral patterning in the eggshell (Table 2 and data not shown). We also observed significant suppression of the oocyte nuclear-morphology defect (Fig. 2g, i). Suppression of the spindle-class defects by mei-41 is not as complete as that by meiW68, raising the possibility that there may be some functional redundancy between mei-41 and a putative Drosophila ATM homologue18, as appears to be the case for yeast MEC1 and TEL1 (reviewed in ref. 20). From our results, we conclude that the pat356

?

?

(grk) (geneX) (Others?) Axial patterning

D/V axis – stage 8-10 Stage 14

Table 2 Suppression of dorsoventral eggshell phenotypes by mei-41

Karyosome morphology

Metaphase-I arrest

Figure 4 Model for the checkpoint-mediated coupling of meiosis to translation of Gurken. Meiotic recombination is initiated by formation of DSBs, a process mediated by mei-W68. Expression of mei-W68 mRNA is detected in regions 2 and 3 (stage-1 egg chamber) of the germarium12. DSBs are repaired by proteins of the recombination-repair pathway, including the proteins encoded by okr and spnB (genes of the spindle class). A mei-41-dependent checkpoint monitors repair of DSBs. In spindle-class mutants, persistence of DSBs activates the mei-41-dependent checkpoint pathway, resulting in modification, perhaps phosphorylation by an unknown kinase, of Vas. Modification of Vas blocks efficient translation of Grk, leading to axial patterning defects in the spindle-class mutants. This block in translation is not absolute, as some Grk protein is produced in all spindle-class mutants. However, if the block in translation is not relieved by inactivation of the checkpoint, dorsoventral pattering, which requires the highest levels of Grk protein, is most strongly affected. In addition, it appears that Vas must also affect another gene (gene X) that is required for normal karyosome formation, as karyosome morphology is abnormal in Vas mutants. It is not clear at what stage of oogenesis double Holliday junctions are produced; however, by analogy with yeast, it seems likely that the appearance of double Holliday junctions will correlate with the loss of synaptonemal complexes, which are no longer visible by stage 7 of oogenesis (reviewed in ref. 30). D/V, dorsoventral; A/P, anteroposterior.

terning defects observed in mutants of the spindle class are caused by the activation of a mei-41-dependent checkpoint pathway in response to the persistence of unrepaired DSBs during meiosis. Vasa is a target of the meiotic checkpoint. These results raise the question of how the mei-41-dependent checkpoint pathway affects accumulation of Grk. One candidate for a downstream target and effector of the mei-41-dependent pathway is the product of the vasa (vas) gene. vas encodes a protein similar to the translation-initiation factor eIF4A, produces mutant phenotypes similar to those observed in okr, spn-B and spn-C mutants, and has been implicated in the translational control of Grk and certain other oocyte-specific proteins4,5,21. However, unlike okr, spn-B and spn-C mutations, mutations in vas are not suppressed by mutations in mei-41 (Table 2). The karyosome phenotype of vas mutants is also not suppressed by mei41 mutations (data not shown). This difference between vas and the spindle-class mutants analysed here indicates that Vas may act downstream of this mei-41-dependent meiotic checkpoint. To address this question more directly, we studied Vas expression in spindle-classmutant backgrounds. At the level of whole-mount antibody staining, Vas does not appear to be affected (data not shown). However, the mobility of Vas, as assessed by SDS polyacrylamide gel electrophoreNATURE CELL BIOLOGY | VOL 1 | OCTOBER 1999 | cellbio.nature.com

articles sis (SDS–PAGE), is altered: Vas migration appears to be retarded in spn-B mutant ovaries as compared with wild-type ovaries or ovaries heterozygous for spn-B (Fig. 3a and data not shown). These results indicate that Vas might be post-translationally regulated by the mei41-dependent checkpoint pathway. In support of this interpretation, the mobility of Vas from ovarian lysates prepared from flies doubly mutant for spn-B and mei-W68 or mei-41 is restored to that observed in wild-type lysates (Fig. 3b). Taken together, these data support a model in which activation of the mei-41-dependent checkpoint pathway occurs in response to the presence of DSBs and leads to the modification of Vas, resulting in the downregulation of its activity and a consequent decrease in Grk translation.

Discussion

Our analysis of the Drosophila spindle-class genes has revealed that progression though meiosis is coupled to specific developmental processes, such as Grk-dependent polarization of the egg chamber along the anteroposterior and dorsoventral axes. At least three of the spindle-class genes appear to function directly in DNA repair, and we have found that mutations in these genes result in the activation of a meiotic DNA-repair checkpoint. Activation of such a meiotic checkpoint in response to the persistence of unrepaired DSBs appears to be a conserved regulatory feature common to yeast, flies and vertebrates. In yeast, activation of the recombination checkpoint downregulates transcription of genes that are targets of Ndt80, including cyclins required for progression through meiosis and a set of sporulation genes required for the morphological changes that normally accompany yeast meiosis22. We propose that, in Drosophila, activation of this checkpoint pathway results in the modification of Vas and downregulation of the translation of Vas targets such as Grk (Fig. 4). In this regard, phosphorylation has been suggested to be a mechanism for downregulating the activity of several translational activators, including two Vas-like proteins, plant eIF4A and vertebrate p68 RNA helicase23–25. Our observations indicate that progression through meiosis can be monitored by assessing repair of DSBs, and progression of meiosis to the stage at which DSBs are repaired can be a prerequisite for translation of key oocyte-specific proteins. In this manner, coupling of translational regulation to progression through meiosis can play a part in the temporal regulation of oocyte development. By delaying translation of certain mRNAs until a particular stage of meiosis is reached, precocious expression of oocyte-specific proteins is prevented. In the early stages of oogenesis, when two or more sister cells initiate meiosis and appear to be following the oocyte developmental programme, it may be crucial to delay expression of oocyte-specific proteins until oocyte identity is firmly established. As homologues of Vas are also specifically expressed in the germ cells of higher eukaryotes26, it will be interesting to see whether regulation of Vas activity by meiotic checkpoints is also conserved in vertebrates. !

Methods Fly strains.

Stocks of mei-P22, c(3)G, mei-W68 (ref. 1) and Df(2R)LL5 mutant flies were gifts from K. McKim12,13. Mutant stocks of mei-41 were obtained from the Bloomington Stock Center and S. Hawley. Various mei41 alleles were used here; D1, D11 and RT1 are believed to be hypomorphic mutations, whereas D3 is amorphic19. The okr RU and AA alleles as well as the spn-B BU and 153 alleles have been described previously1. Complementation tests were performed between spn-C094 and the D1, D2 and D4 alleles of mus301, as well as between spn-C422 and mus301D4. All of these combinations fail to complement. The relative strengths of these spn-C/mus301 alleles in trans to a Df allele that uncovers spn-C — Df(3L)pblX1 — are, in order of increasing severity, 094
Competitive PCR. Evalulation of the relative levels of grk mRNA in wild-type (Ore-R) as compared with in spn-B, spn-C (mus301) and okr mutant ovaries was done as in ref. 5. Primers designed to amplify a region of actin5C mRNA were used to normalize the samples. We used the BU and 153 alleles of spn-B (153 is shown in Fig. 1), the D2 allele of spn-C (mus301) in trans to a deficiency that uncovers the spn-C region, and the RU allele of okr.

NATURE CELL BIOLOGY | VOL 1 | OCTOBER 1999 | cellbio.nature.com

Fluorescent images and confocal microscopy OliGreen (Molecular probes) and Grk staining were studied with a Bio-Rad MRC 600 confocal microscope as described1, with the exceptions that egg chambers were fixed in 2% paraformaldehyde and that anti-Grk monoclonal antibodies ID12 and IF12 were used. For detection, Alexa-568-conjugated anti-mouse IgG (Molecular Probes) was used at a dilution of 1:1,000.

Western blotting. Ovarian extracts were prepared as described28. Proteins were resolved by SDS–PAGE on 12% low Bis gels29 at 200 V for 13.5 h. Rabbit-anti-Vas serum was used at a dilution of 1:500. Biotinylated-anti-rabbit serum (Jackson) was used at 1:1,000. Streptavidin-conjugated horseradish peroxidase (Amersham) was used at 1:1,000 in combination with the enhanced chemiluminescence plus kit (Amersham) for immunodetection. RECEIVED 1 JUNE 1999; REVISED 21 JULY 1999; ACCEPTED 12 AUGUST 1999; PUBLISHED 9 SEPTEMBER 1999.

1. Ghabrial, A., Ray, R. P. & Schupbach, T. okra and spindle-B encode components of the RAD52 DNA repair pathway and affect meiosis and patterning in Drosophila oogenesis. Genes Dev. 12, 2711–2723 (1998). 2. Gonzalez-Reyes, A., Elliott, H. & St Johnston, D. Oocyte determination and the origin of polarity in Drosophila: the role of the spindle genes. Development 124, 4927–4937 (1997). 3. Ray, R. P. & Schüpbach, T. Intercellular signaling and the polarization of body axes during Drosophila oogenesis. Genes Dev. 10, 1711–1723 (1996). 4. Styhler, S. et al. vasa is required for Gurken accumulation in the oocyte, and is involved in oocyte differentiation and germline cyst development. Development 125, 1569–1578 (1998). 5. Tomancak, P., Guichet, A., Zavorszky, P. & Ephrussi, A. Oocyte polarity depends on regulation of gurken by vasa. Development 125, 1723–1732 (1998). 6. Kooistra, R. et al. The Drosophila melanogaster RAD54 homolog, DmRAD54, is involved in the repair of radiation damage and recombination. Mol. Cell. Biol. 17, 6097–6104 (1997). 7. Boyd, J. B., Golino, M. D., Shaw, K. E., Osgood, C. J. & Green, M. M. Third-chromosome mutagensensitive mutants of Drosophila melanogaster. Genetics 97, 607–623 (1981). 8. Petukhova, G., Stratton, S. & Sung, P. Catalysis of homologous DNA pairing by yeast Rad51 and Rad54 proteins. Nature 393, 91–94 (1998). 9. Roeder, G. S. Meiotic chromosomes: it takes two to tango. Genes Dev. 11, 2600–2621 (1997). 10. Keeney, S., Giroux, C. N. & Kleckner, N. Meiosis-specific DNA double-strand breaks are catalyzed by Spo11, a member of a widely conserved protein family. Cell 88, 375–384 (1997). 11. Cao, L., Alani, E. & Kleckner, N. A pathway for generation and processing of double-strand breaks during meiotic recombination in S. cerevisiae. Cell 61, 1089–1101 (1990). 12. McKim, K. S. & Hayashi-Hagihara, A. mei-W68 in Drosophila melanogaster encodes a Spo11 homolog: evidence that the mechanism for initiating meiotic recombination is conserved. Genes Dev. 12, 2932–2942 (1998). 13. McKim, K. S. et al. Meiotic synapsis in the absence of recombination. Science 279, 876–878 (1998). 14. Gowen, J. W. Meiosis as a genetic character in Drosophila melanogaster. J. Exp. Zool. 65, 83–106 (1933). 15. Bishop, D. K., Park, D., Xu, L. & Kleckner, N. DMC1: a meiosis-specific yeast homolog of E. coli recA required for recombination synaptonemal complex formation and cell cycle progression. Cell 69, 439–456 (1992). 16. Pittman, D. L. et al. Meiotic prophase arrest with failure of chromosome synapsis in mice deficient for Dmc1, a germline-specific RecA homolog. Mol. Cell 1, 697–705 (1998). 17. Lydall, D., Nikolsky, Y., Bishop, D. K. & Weinert, T. A meiotic recombination checkpoint controlled by mitotic checkpoint genes. Nature 383, 840–843 (1996). 18. Sekelsky, J. J., Burtis, K. C. & Hawley, R. S. Damage control: the pleiotropy of DNA repair genes in Drosophila melanogaster. Genetics 148, 1587–1598 (1998). 19. Sibon, O. C. M., Laurencon, A., Hawley, R. S. & Theurkauf, W. E. The Drosophila ATM homologue Mei-41 has an essential checkpoint function at the midblastula transition. Curr. Biol. 9, 302–312 (1999). 20. Bentley, N. J. & Carr, A. M. DNA structure-dependent checkpoints in model systems. Biol. Chem. 378, 1267–1274 (1997). 21. Lasko, P. F. & Ashburner, M. The product of the Drosophila gene vasa is very similar to eukaryotic initiation factor-4A. Nature 335, 611–617 (1988). 22. Chu, S. & Herskowitz, I. Gametogenesis in yeast is regulated by a transcriptional cascade dependent on Ndt80. Mol. Cell 1, 685–696 (1998). 23. Buelt, M. K., Glidden, B. J. & Storm, D. R. Regulation of p68 RNA helicase by calmodulin and protein kinase C. J. Biol. Chem. 269, 29367–29370 (1994). 24. Gallie, D. R. et al. The phosphorylation state of translation initiation factors is regulated developmentally and following heat shock in wheat. J. Biol. Chem. 272, 1046–1053 (1997). 25. Webster, C. et al. Hypoxia enhances phosphorylation of eukaryotic initiation factor 4A in maize root tips. J. Biol. Chem. 266, 23341–23346 (1991). 26. Olsen, L. C., Aasland, R. & Fjose, A. A vasa-like gene in zebrafish identifies putative primordial germ cells. Mech. Dev. 66, 95–105 (1997). 27. Lindsley, D. L. & Zimm, G. G. The Genome of Drosophila melanogaster (Academic, New York, 1992). 28. Gavis, E. R. & Lehmann, R. Translational regulation of nanos by RNA localization. Nature 369, 315– 318 (1994). 29. Whalen, A. M. & Steward, R. Dissociation of the Dorsal-Cactus complex and phosphorylation of the Dorsal protein correlate with the nuclear localization of Dorsal. J. Cell Biol. 3, 523–534 (1993). 30. Spradling, A. C. in The Development of Drosophila melanogaster (eds Bate, M. & Martinez Arias, A.) 1–70 (Cold Spring Harb. Lab. Press, Cold Spring Harbor, 1993). ACKNOWLEDGEMENTS We thank K. McKim, J. Sekelsky, S. Hawley, S. Wayson and R. Ray for mutant stocks and helpful discussions; C. VanBuskirk for sharing anti-Grk monoclonal antibodies; I. Clark for Vasa reagents and advice on Vasa westerns; G. Shanower, G. Deshpande and P. Schedl for their advice; and E. Wieschaus, L. Nilson, C. VanBuskirk and A. Norvell for comments on the manuscript. This work was supported by the US Public Health Service grant PO1 CA 41086 and the HHMI. Correspondence and requests for materials should be addressed to T.S.

357