Horka^D, a Chromosome Instability-Causing Mutation in Drosophila, Is a Dominant-Negative Allele of lodestar

Abstract

Correct segregation of chromosomes is particularly challenging during the rapid nuclear divisions of early embryogenesis. This process is disrupted by Horka^D, a dominant-negative mutation in Drosophila melanogaster that causes female sterility due to chromosome tangling and nondisjunction during oogenesis and early embryogenesis. Horka^D also renders chromosomes unstable during spermatogenesis, which leads to the formation of diplo//haplo mosaics, including the gynandromorphs. Complete loss of gene function brings about maternal-effect lethality: embryos of the females without the Horka^D-identified gene perish due to disrupted centrosome function, defective spindle assembly, formation of chromatin bridges, and abnormal chromosome segregation during the cleavage divisions. These defects are indicators of mitotic catastrophe and suggest that the gene product acts during the meiotic and the cleavage divisions, an idea that is supported by the observation that germ-line chimeras exhibit excessive germ-line and cleavage function. The gene affected by the Horka^D mutation is lodestar, a member of the helicase-related genes. The Horka^D mutation results in replacement of Ala777 with Thr, which we suggest causes chromosome instability by increasing the affinity of Lodestar for chromatin.

MOST of the proteins required in early embryogenesis are maternally provided; it is generally agreed that little if any zygotic gene expression occurs during the onset of embryogenesis (Derenzo and Seydoux 2004; Tadros and Lipshitz 2005). To dissect the commencement of embryogenesis in Drosophila melanogaster, we isolated dominant female-sterile (Fs) mutants (Erdelyi and Szabad 1989; Szabad et al. 1989) and focused our attention on those that terminate embryogenesis at or shortly after fertilization.

Horka^D is one such Fs mutation (Erdelyi and Szabad 1989). It is a gain-of-function mutation (Erdelyi and Szabad 1989) that results in chromosome nondisjunction and renders chromosomes unstable during spermatogenesis, causing them to be lost in the resulting zygotes (Szabad et al. 1995). Loss of the chromosomes leads to the formation of diplo//haplo mosaics, including XX//X0, female//male mosaics, and gynandromorphs (Szabad et al. 1995). (X represents chromosomes derived from the Horka^D males.) In fact, Horka^D has been used as a “tool” to generate genetic mosaics (Szabad and Nothiger 1992; Zallen and Wieschaus 2004; Villanyi et al. 2008).

To determine the function of the gene carrying the Horka^D mutation, we first mapped Horka^D by screening for duplications that can ameliorate the Horka^D mutant phenotype in embryos of the Horka^D/+/+ females. This revealed the dominant-negative (antimorphic) nature of the mutation. We generated horka^rvP P-element-induced alleles (hereafter called pseudorevertants) that no longer exhibit the dominant mutant phenotype and used them to map and then isolate the gene. We discovered that Horka^D is an allele of lodestar (lds), which encodes a member of the Snf2 family of the helicase-related genes (Girdham and Glover 1991; Liu et al. 1998; Flaus et al. 2006). Our results suggest that the lodestar (LDS) protein is involved in progression from metaphase to anaphase of the cell cycle. We propose that the lodestar protein altered by Horka^D disturbs chromatin organization and segregation and renders chromosomes unstable. It appears thus that the LDS protein is one of the many components engaged in maintaining genome integrity (Takada et al. 2003; Allard et al. 2004; Musacchio and Salmon 2007).

MATERIALS AND METHODS

Horka^D, horka^rv, and Horka^RR alleles:

Horka^D was induced by EMS on an isogenic third chromosome labeled with the mwh and the e recessive marker mutations (Erdelyi and Szabad 1989). For an explanation of the genetic symbols, see FlyBase at http://flybase.bio.indiana.edu. The horka^rv revertant alleles were generated through second mutagenesis of Horka^D: the horka^rvE1 allele by EMS (Erdelyi and Szabad 1989) and the horka^rvP alleles through mutagenesis with the normal P elements. For induction of the horka^rvP alleles, dysgenic Horka^D/TM3, Sb Ser males were mated with TM3, Ser/TM1, Me virgin females. The P elements were hopping in these males and might have become inserted into the Horka^D allele. (The dysgenic Horka^D/TM3, Sb Ser males were generated by crossing M cytotype TM3, Sb Ser/TM6B, Tb females with P cytotype Horka^D/TM3, Sb Ser males. The latter males resulted from a cross between P cytotype CxD/TM3, Sb Ser females and Horka^D/TM3, Sb Ser males.) Since the TM3, Sb Ser/TM3, Ser and the TM3, Sb Ser/TM1, Me combinations are lethal, only the Horka^D/TM3, Ser and the Horka^D/TM1, Me offspring survive. The resulting females, who mated with the sibling males, were screened for offspring production. Only the horka^rvP/TM3, Ser and the horka^rvP/TM1, Me females give rise to progeny, allowing a direct selection of the horka^rvP phenotypically revertant alleles. (To avoid the isolation of clusters of the horka^rvP alleles, groups of 10 dysgenic males were mated with TM3, Ser/TM1, Me females and the descendants from the parallel crosses were screened separately.)

The P-element insertion sites in the horka^rvP alleles were determined by standard in situ hybridization on salivary gland chromosomes, using DIG-labeled P-element DNA probe.

To remobilize the P elements in the horka^rvP revertants and isolate Horka^RR alleles (revertant alleles of the horka^rvP revertants), we constructed horka^rvP/TM3, Δ2-3 females and males. The Horka^RR originated most likely through precise excision of the P element from the horka^rvP alleles. The Horka^RR alleles, which behaved as Horka^D, were used in in situ hybridization studies on salivary gland chromosomes.

The chromosome destabilizing effect of Horka^D and the horka^rvP alleles was analyzed in outcrosses with y v f mal females and measured through the frequency of XX//X0, female//male mosaics among the descending XX zygotes (cf. Szabad et al. 1995).

Drosophila cultures used in the study were kept at 25°.

The Horka^D/Dp(3;3) combinations:

Horka^D was mapped to the right arm of the third chromosome (Erdelyi and Szabad 1989). To determine the approximate location and the nature of Horka^D (whether it is antimorphic or neomorphic), we constructed Horka^D/Dp(3;3) females and males by crossing Dp(3;3)/TM3 females with Horka^D/TM3, Sb Ser males. Dp(3;3) stands for 18 tandem duplications, which cover—bit by bit—the right arm of the third chromosome. The resulting Horka^D/Dp(3;3) females were mated with wild-type males and the fate of their resulting embryos was monitored. Males were mated with y v f mal females and the subsequent generation was screened for XX//X0 mosaics.

Localizing the horka^rv alleles and complementation analysis:

To locate the horka^rv alleles and to determine the loss-of-function mutant phenotype, we combined the horka^rv alleles (as well as Horka^D) with Df(3R) deficiencies and analyzed the horka^rv/− (and the Horka^D/−) flies. (The − symbol stands for either of the deficiencies that remove the Horka^D-identified locus.) The studied Horka^D/− and the horka^rv/− hemizygotes were produced by crossing Df(3R)dsx¹⁵/TM6B, Tb females with horka^rv/TM6B, Tb or with Horka^D/TM6B, Tb males.

To determine whether the horka^rv alleles identify a gene with already existing mutant alleles, we carried out complementation analyses between horka^r and mutant alleles of the nearby genes, lds, dsx, and CG10445 (see Figure 3). (Mutant alleles of the CG10445 gene were generated in our laboratory; I. Belecz and J. Szabad, unpublished observations.)

Figure 3.—

Organization of the region around the lodestar gene in the 84E5 cytological region. The lds gene encodes the formation of two mRNAs that differ in the last ∼500 nucleotides. Stippled boxes correspond to sequences that encode the 5′ and the 3′ untranslated regions of the lds mRNAs, and open and solid boxes represent introns and exons, respectively. The P-element insertion sites in horka^rvP3, horka^rvP9, and horka^rvP2 are labeled and also the position of the Horka^D mutation (). The shaded lines represent different transgene types.

Characterization of mutant phenotypes:

To describe the Horka^D- and the horka^rv-associated mutant phenotypes, ovaries, testes, and eggs/embryos of Horka^D/+; Horka^D/− and horka^rvP2/− females and males were dissected and fixed according to González and Glover (1993). The stage 14 oocytes were immunostained according to Tavosanis et al. (1997). The eggs and the embryos were prepared as follows: the chorion was removed by Clorox, the dechorionated embryos were fixed in a 1:1 mixture of 4% paraformaldehyde:heptane or in a 1:1 mixture of methanol:heptane, and the vitelline membrane was removed subsequently by agitation in a mixture of heptane and methanol. To block nonspecific staining, the embryos were incubated in 1% BSA (Sigma, St. Louis) in PBST for 90 min at room temperature.

For immunological detection of the microtubules, we used the DM1A monoclonal anti-α-tubulin antibody (1:1000, overnight at 4°; T6199, Sigma). The centrosomes were detected using an anti-centrosomin antibody (Heuer et al. 1995). The LDS protein was detected by polyclonal anti-LDS rabbit antibody raised against the almost complete LDS protein, a generous gift from David Glover's laboratory (Girdham and Glover 1991). The anti-LDS antibody was present in the serum from which the nonspecific components were depleted through preincubation of the serum in dechorionated and heptane-permeabilized eggs of horka^rvP2/− females, which do not contain LDS protein. The anti-LDS antibody was applied at a 1:200 dilution in 1% BSA in PBST. The embryos were incubated in secondary antibodies for 3 hr at room temperature or overnight at 4°. The secondary antibodies were either anti-mouse or anti-rabbit IgG (Sigma) and were labeled with FITC, Texas-Red, or Alexa Fluor-633. To detect DNA, the embryos were stained with DAPI following incubation with the secondary antibody. After several rinses in PBST, the embryos and the testes were mounted in Aqua PolyMount (Polysciences, Warrington, PA). The immunostained preparations were analyzed either in an Olympus IX71 fluorescent microscope with a cooled CCD camera or through optical sections collected in an Olympus FV1000 confocal microscope.

We also prepared and analyzed cuticles of the dead embryos as described in Wieschaus and Nusslein-Volhard (1989).

Cytoplasm injections:

To analyze the effect of Horka^D on the cleavage divisions, we injected ∼300 pl of cytoplasm (∼3% of the total egg volume) from eggs of wild-type (as the control) and Horka^D/+ females into the posterior region of embryos in which the microtubules were highlighted by Jupiter-GFP and the nuclei by histone-RFP (Karpova et al. 2006; Schuh et al. 2007). The donor embryos were a maximum of 30 min old and the injected embryos were in the 9th–11th cleavage cycle of embryogenesis. Effect of the injected cytoplasm was followed in time through a series of optical sections generated in an Olympus FV1000 confocal microscope. The injections were carried out at 25°.

Germ-line chimeras:

To determine whether the Horka^D- and the horka^rvP2/− -related defects originate from altered function of the germ line and/or the soma, we constructed different types of germ-line chimeras through the transplantation of pole cells, embryonic precursor cells of the future germ line. Tables 2 and 3 list the crosses from which the donor and the host embryos originated. Pole cells were collected from single blastoderm-stage donor embryos and transplanted into two to three host blastoderm embryos. While pole cells do not develop in the embryos of the tropomyosin-II^gs (tmII^gs) homozygous females, the somatic cells function normally (Erdelyi et al. 1995). Fs(1)K1237 (also known as ovo^D1) is an X-linked dominant female-sterile mutation (Komitopoulou et al. 1983; Perrimon 1984). Although the Fs(1)K1237/+ host females do not produce eggs of their own, their soma provides a normal environment for development of the received female pole cells. Pole cells of y v f mal embryos were transplanted into Horka^D/+ and horka^rvP2/− host embryos, and the developing adults were analyzed for the presence of the implanted y v f mal germ-line cells. The flies that developed following pole cell transplantation were mated with appropriate partners, as described in Tables 2 and 3, and tested for germ-line chimerism.

TABLE 2.

Features of the Horka^D/+ germ-line chimeras

Cross to produce the donor embryos			Stock to produce the donor embryos
mwh e/mwh e ♀♀ × HorkaD/TM3 ♂♂			y v f mal
↓			↓
Cross to produce the host embryos			Cross to produce the host embryos
tmIIgs/tmIIgs ♀♀ × tmIIgs/TM6 ♂♂			mwh e/mwh e ♀♀ × HorkaD/TM3 ♂♂
Genotype of the transplanted pole cells	Germ-line chimera		Genotype of the host embryos	Germ-line chimera
		Femalea		Maleb		Femalec	Malec
mwh e/HorkaD	3	3	mwh e/HorkaD	4	2d
mwh e/TM3	8	4	mwh e/TM3	5	2

^aThe females were mated with mwh e/mwh e males.

^bThe males were mated with y v f mal females.

^cMated with y v f mal partner.

^dMany more offspring originated from the y v f mal than from their own Horka^D/+ germ-line cells.

TABLE 3.

Features of the horka^rvP2/− germ-line chimeras

Cross to produce the donor embryos		Stock to produce the donor embryos
horkarvP2/TM6B ♀♀ × Df(3R)dsx15/TM3 ♂♂		y v f mal
↓		↓
Cross to produce the host embryos		Cross to produce the host embryos
w/w ♀♀ × Fs(1)K1237/Y ♂♂		horkarvP2/TM6B ♀♀ × Df(3R)dsx15/TM3 ♂♂
Genotype of the transplanted pole cells	Germ-line chimera	Genotype of the host embryos	Germ-line chimera
						Femalea	Malea
horkarvP2/TM3	8	horkarvP2/TM3	2	2
Df(3R)dsx15/TM6B	5	Df(3R)dsx15/TM6B	3	4
TM3/TM6B	1	TM3/TM6B	1	2
horkarvP2/Df(3R)dsx15	3	horkarvP2/Df(3R)dsx15	1	1

Arrows symbolize the direction of pole cell transplantations. Horka^D was induced by EMS on an mwh- and e-labeled isogenic chromosome (Erdelyi and Szabad 1989).

^aThe chimeras produced y v f mal offspring following test crosses with y v f mal partners.

Inverse PCR:

To clone the Horka^D-identified gene, we used the inverse PCR technique and amplified DNA sequences flanking the P elements in three of the horka^rvP alleles. Briefly, we isolated DNA from horka^rvP-carrying males and digested the DNA with HinPI or with MspI. The digested genomic DNA was ligated overnight at 4°, ethanol precipitated, and resuspended in distilled water. Two PCR reactions were conducted next. In the first reaction, the outward primers were designed on the basis of the terminal sequences of the P-element adjacent to the cut site. (The primers are described in supplemental Table 2.) Because the first PCR did not yield sufficient amounts of DNA for sequencing, a second, so-called nested PCR reaction was conducted using primers complementary to slightly more interior sequences in the P element. (See supplemental Table 2.) Products from the second PCR reactions were isolated, purified, and sequenced in an IBI automated sequenator on both strands. The resulting sequence information allowed us to precisely position the P-element insertion sites on the Drosophila genome sequence (Adams et al. 2000).

Molecular cloning and sequencing of Horka^D:

DNA of Horka^D/– and mwh e (as the control) males served as a template in a set of PCR reactions to produce DNA fragments for sequencing. The PCR primers were designed on the basis of the lds gene sequence (EMBL nucleotide sequence database, accession no. X62629). Sequencing of the PCR products was carried out in an IBI sequenator on both strands.

The Horka⁺ (TG⁺) and the Horka^D (TG^HD) transgenes:

To characterize the Horka^D-identified gene, we generated a Horka⁺ transgene (TG⁺) in which a 5.1-kb genomic segment included both the regulatory and the structural parts of the lds gene (see Figure 3). The 5107-bp genomic sequence was cloned into the CaSpeR vector with the mini-white marker gene and a germ-line transformant transgenic line was generated on a w¹¹¹⁸ background by standard procedures. The TG⁺ transgene became inserted into the second chromosome. The TG⁺ transgene was combined, in appropriate genetic crosses, with the Horka^D, the horka^rv, and the lds mutant alleles to determine whether the TG⁺ transgene can overcome the mutant phenotypes associated with the Horka^D and the horka^rv alleles.

To generate transgenes that carry Horka^D (the TG^HD transgenes), we PCR amplified a 5107- and a 5499-bp genomic segment that included the promoter and the structural parts of the Horka^D allele (see Figure 3). The DNA was isolated from Horka^D/Df(3R)dsx¹⁵ males. The two transgene types correspond to the two lds mRNAs that differ by ∼500 nucleotides in their 3′-UTR (Girdham and Glover 1991; see Figure 3). Stable germ-line transformant lines were generated through standard procedures.

RESULTS

Horka^D disrupts the meiotic and the early cleavage divisions:

Although the Horka^D/+ females deposit normal numbers of normal-looking eggs (fertilized as in wild type), cleavage divisions do not commence in >90% of the eggs. Moreover, when cleavage divisions are seen, only ∼12 scattered chromosomes appear, along with unusual microtubule bundles (Figure 1). As expected, cuticle fragments, indicators of development beyond the blastoderm stage, never form inside the eggs of the Horka^D/+ females (Table 1). Abnormal segregation of the chromosomes is already apparent during both the first and the second meiotic divisions in egg primordia of the Horka^D/+ females (Figure 1). The mutant phenotypes suggest involvement of the Horka^D-identified normal gene product in chromosome organization, stability, and/or segregation.

Figure 1.—

Meiotic and cleavage divisions in the egg primordia and in eggs of wild-type, Horka^D/+, and horka^rvP2/− females. In the optical sections the microtubules appear in green, the centrosomes and the spindle pole bodies in red, and the DNA in blue. Detachment of one of the spindles (dashed circle) is a typical feature of the second meiotic division in the Horka^D/+ females. Although the meiotic divisions proceed as in wild type in ∼70% of the cases, abnormal meiotic spindles develop in a number of egg primordia in the horka^rvP2/− females. Note that most centrosomes cannot nucleate astral microtubules and several of the spindles are abnormal in embryos of the horka^rvP2/− females. Bars, 10 μm.

TABLE 1.

Features of the Horka^D- and the horka^rvP2-carrying females and males

	Analysis of the females					Analysis of the male offspringc
Test perioda	Dead embryos with cuticle (%)	Rate of offspring productionb	XX//X0 mosaic
			Genotype	Tested			Offspring		XX	Total	%
HorkaD/+	851	16.3	0	0	—	4304	432	9.1
HorkaD/Dp(3;3)d	1704	8.8	0	0	—	—	—	2.1–28.8
HorkaD/Dp(3;3)Antprv8	261	18.0	100	3	6.4 × 10−4	116	20	14.7
HorkaD/Df(3R)dsx15	161	19.5	0	0	—	178	14	7.3
HorkaD/lds98.1	310	20.7	0	0	—	276	2	0.7
TG+; HorkaD/+	258	19.2	93	0	—	363	14	3.7
+/+; TGHD5.1	147	16.7	28	0	—	167	13	7.2
+/+; TGHD5.5	188	15.5	25	0	—	246	8	3.2
horkarvP2/Df(3R)dsx15	180	15.2	20.6	—	—	194	0	—
horkarvP2/+	11	7.0	3.4	2695	35.0	3433	0	—
horkarvP2/lds98.1	85	12.3	21.0	—	—	298	0	—
TG+; horkarvP2/Df(3R)dsx15	5	7.0	11.0	087	31.1	—	—	—
TG+; horkarvP2/lds98.1	7	7.0	8	649	33.7	—	—	—

horka^rvP2 is a functionally null allele and lds^98.1 is a lodestar null allele (Girdham and Glover 1991).

^aAverage test period per female (days).

^bOffspring/(female × day).

^cThe males were mated with y v f mal females (XX) and the XX offspring flies were screened for XX//X0 mosaics.

^dPooled data from 17 Dp(3;3) tandem duplications with the exception of Dp(3;3)Antp^rv8.

Horka^D has been reported to be a gain-of-function mutation (Erdelyi and Szabad 1989). We have now confirmed this observation by cytoplasm injection studies. When cytoplasm taken from newly deposited eggs of the Horka^D/+ females was injected into horka⁺ embryos in which the chromosomes were highlighted by RFP-tagged histones and the microtubules by GFP-tagged tubulins, it induced chromosome tangling during ana- and telophase, the formation of chromatin bridges, abnormally shaped and positioned nuclei (which usually drop inside the egg cytoplasm during the upcoming cleavage mitosis), and free centrosomes. (See Videos 1 and 2 in the supplemental material.) Toxicity of the Horka^D-derived egg cytoplasm is best illustrated by the fact that not a single embryo survived the cytoplasm injections. (Injection of wild-type egg cytoplasm did not alter progression of the cleavage cycles, and larvae hatched from almost all of the injected embryos.)

Horka^D resides between 84D5–8 and 85F5–8:

Horka^D has been mapped to the right arm of the third chromosome (Erdelyi and Szabad 1989). To more accurately locate Horka^D, we generated a series of Horka^D/Dp(3;3) flies and analyzed the embryos of the females and searched for XX//X0 mosaics among the XX offspring of the males. If Horka^D is a dominant-negative mutation, (i) a less severe mutant phenotype was expected to develop inside eggs of the Horka^D/Dp(3;3)⁺ females (compared to the Horka^D/+ control) and (ii) reduced frequency of the XX//X0 mosaics was expected to appear in the offspring of the Horka^D/Dp(3;3)⁺ males. We used 18 Dp(3;3) tandem duplications that, in aggregate, fully covered the entire right arm of the third chromosome. Of these duplications tested, only Dp(3;3)Antp^rv8 ameliorated the Horka^D imposed defects, such that embryogenesis inside eggs of the Horka^D/Dp(3;3)Antp^rv8 females proceeded well beyond the initial steps. Not only did cuticle fragments form in almost 100% of the eggs, but also three offspring were produced by the Horka^D/Dp(3;3)Antp^rv8 females (Table 1). Thus the results of mapping located Horka^D within the 84D5–8 and 85F5–8 cytological interval (Figure 2). Moreover, the ability of Dp(3;3)Antp^rv8 to ameliorate the dominant-negative nature of Horka^D argues that the mutant and normal gene products participate in the same process and the mutant gene product is truly antimorphic (i.e., it impedes function of the normal counterpart).

Figure 2.—

Duplication and deficiency mapping of the Horka^D and the horka^rvP2 mutations. The thick bar represents the Dp(3;3)Antp^rv8 tandem duplication. Missing sections in the Df(3R) deficiencies illustrate the eliminated chromosome segments. The deficiencies located the horka^rv-identified locus between 84E1 and 84E8.

We cannot discern whether the Horka^D-related dominant paternal effect is also of dominant-negative nature because the frequency of the XX//X0 mosaics varied between 2.1 and 28.8% in the XX offspring of the Horka^D/Dp(3;3) males. The variation in the XX//X0 mosaic frequencies is most likely related to the different genetic backgrounds of the Horka^D/Dp(3;3) males (Szabad et al. 1995).

The horka^{pseudorevertant} (horka^rv) alleles:

To learn the function of the gene carrying the Horka^D mutation we induced horka^rvP pseudorevertant alleles through mutagenesis of Horka^D with a normal P element. Of the 15,600 P-element-mutagenized females tested, 9 independent ones were fertile and gave rise to one horka^rvP pseudorevertant allele each. All the horka^rv alleles are lethal both in homozygotes and in trans-heterozygotes due to one or more second-site lethal mutations induced during EMS induction of Horka^D (cf. Erdelyi and Szabad 1989).

To characterize the horka^rvP alleles, we crossed horka^rvP/TM3, Sb Ser males from each of the nine horka^rvP lines with y v f mal females and searched the offspring for XX//X0 mosaics (see Table 1). Mosaics appeared (though with very low frequencies) among the offspring of three of the nine horka^rvP pseudorevertants (horka^rvP5, horka^rvP6, and horka^rvP8), suggesting that these alleles retain some feature of Horka^D. Their incomplete loss of the Horka^D phenotype is also shown by the strongly reduced fertility of the heterozygous females. Mosaics did not appear in the offspring of the males that were heterozygous for the other six horka^rvP alleles and larvae hatched from the vast majority of the eggs deposited by the heterozygous females, indicating the loss-of-function nature of six of the horka^rvP mutations. One of the alleles, horka^rvP2, is a complete loss-of-function mutation (Table 1 and see below). The concurrent loss of dominant female sterility and dominant paternal effect in six of the nine horka^rvP alleles shows that the Horka^D-related dominant mutant phenotypes stem from the same mutation.

The horka^r mutations reside within the 84E1–84E8 cytological interval:

The horka^rv alleles (and also Horka^D) were combined with deficiencies that remove well-defined regions around the 84E cytological region (Figure 2). The horka^rv/– hemizygous combinations are viable and the flies develop with the expected frequencies. (The – symbol stands for either of the deficiencies that remove the horka^rv identified locus.) The horka^rv/– hemizygous females either are completely sterile (horka^rvP2/–; Table 1) or possess reduced fertility: progeny develop from 4–21% of the zygotes in all the other horka^rv/– combinations. The deficiencies located the horka^rv-identified locus within the 84E1–E8 cytological region (Figure 2).

The fertility of horka^rvP2/– males is also very strongly reduced (see supplemental Table 2). The cross in which several hundred horka^rvP2/– males were mated with several hundred y v f mal females yielded only few offspring, none of which was XX//X0 mosaic (Table 1).

The Horka^D/– flies are also viable and emerge with the expected frequency. The females deposit normal numbers of normal-looking eggs in which, although normally fertilized, embryogenesis never commences. Fertility of the Horka^D/– males is also strongly reduced (supplemental Table 2). However, a few offspring derived from a cross between several hundred Horka^D/– males and y v f mal females and 6.7% (14/192) of the XX offspring flies were XX//X0 mosaics (Table 1).

It appears that reduced fertility of the Horka^D also exhibits a dominant-negative effect on male fertility, as evidenced by the observation that, when sired by Horka^D/–, Horka^D/+, or Horka^D/Dp(3;3)Antp^rv8 males, 92, 71, and 59% of the embryos perished during embryogenesis (supplemental Table 2). Because there was no sperm in 98.5% (446/453) of the eggs in which embryogenesis did not commence, the reduced fertility of the Horka^D-carrying males is most likely the consequence of abnormal spermatogenesis. Remarkably, the egg production rate of the partner y v f mal females was not significantly different from the control (supplemental Table 2) and thus Horka^D and horka^rvP2 do not seem to affect other fertility-related features than sperm production, suggesting that Horka^D has little if any effect on the somatic cells (cf. Liu and Kubli 2003).

In situ hybridizations confirm that the Horka^D-identified gene resides in 84E:

To locate the gene carring the Horka^D mutation we probed salivary gland chromosomes in the nine horka^rvP revertants with labeled P-element DNA. There were three to six P-element insertions in the right arm of the third chromosome in the different horka^rvP alleles. The only common P-element insertion site appeared in 84E, suggesting that the gene resides in 84E.

The P elements of the horka^rvP7 and the horka^rvP9 alleles were successfully remobilized. As a result, two Horka^RR alleles (revertant alleles of the horka^rvP mutations) emerged from the 135 chromosomes tested. The Horka^RR alleles behaved as Horka^D: Horka^RR/+ females are sterile and embryos perish inside their eggs, essentially as described for the Horka^D/+ females. Among the progeny of the Horka^RR/+ males and y v f mal females, 13.9% (14/85) and 13.0% (25/164) of the XX zygotes developed as XX//X0 mosaics in the two Horka^RR alleles. More important, the P-element insertions in 84E were absent in the Horka^RR alleles, suggesting that the gene carrying the Horka^D mutation indeed resides at 84E. The Horka^RR alleles underline the common origin of the Horka^D-related dominant defects.

Horka^D and its revertant alleles identify the lodestar gene:

We exploited the P elements in horka^rvP2, horka^rvP3, and horka^rvP9 (each with as few as three P elements inserted into 3R) to identify the gene carrying Horka^D. We amplified sequences adjacent to the P elements in an inverse PCR. The DNA sequence of the PCR products was determined, and we analyzed only those originating from 84E. The P element resides in the leader sequence of the lds gene in horka^rvP3 and in horka^rvP9 (Figure 3 and supplemental Figure 3). In horka^rvP2, the P element is inserted between nucleotides 3,746,261 (G) and 3,746,262 (A) in the first exon of the open reading frame of the lds gene. On the basis of this mutant lesion and the observation that the LDS protein is also missing from ovaries of the horka^rvP2/− females (data not shown), we conclude that horka^rvP2 is a null allele of the lds gene.

The positions of the P elements in the horka^rvP alleles identify the lds gene (Girdham and Glover 1991). Indeed, the horka^rv and the lds alleles do not complement, so we conclude that Horka^D is a dominant-negative lds allele (Table 1). The horka^rvP2/lds^98.1 and the horka^rvP2/lds^298.8 combinations are female sterile and are as strong as the horka^rvP2/− condition. The lds^98.1 and the lds^298.8 alleles are complete loss-of-function alleles as there is no LDS protein in ovaries of the lds^98.1/− and the lds^298.8/− hemizygous females (Girdham and Glover 1991). Although embryogenesis proceeds beyond the blastoderm stage in ∼60% of the eggs of the horka^rvP2/−, the horka^rvP2/lds^98.1, and the horka^rvP2/lds^298.8 females and fragments of cuticles appear in ∼21% of their eggs, larvae never hatch. The females are semisterile in all the further horka^rv/lds combinations.

We crossed several hundred horka^rvP2/lds^98.1 males, which are almost completely sterile, with several hundred y v f mal females. None of the recovered 298 XX offspring were XX//X0 mosaic (Table 1). However, XX//X0 mosaics appeared among offspring of the Horka^D/lds^98.1 males (Table 1).

An lds⁺-bearing transgene (denoted TG⁺) rescues loss-of-function horka mutants:

To confirm that Horka^D and horka^rv alleles indeed identify the lds gene we generated a stable transgenic line (TG⁺, inserted into a second chromosome) that covers a 5.1-kb genomic sequence and includes the normal lds gene (except the last 500 bp; Figure 3). Although the TG⁺; Horka^D/+ females are sterile, the effects of Horka^D are ameliorated and cuticle fragments develop inside 93% of their eggs (Table 1). The TG⁺ transgene overcomes the sterility of the horka^rvP2/−, horka^rvP2/lds, and lds^98.1/lds^298.8 females (Table 1). In the presence of TG⁺, fertility of the horka^rvP2/− and the horka^rvP2/lds^98.1 males is essentially as in wild type (see supplemental Table 2). Evidently, Horka^D and horka^rv are alleles of the lds gene.

Horka^D originated through a transition:

The Horka^D mutation is a single-nucleotide change G²⁴²⁴ → A, resulting in the replacement of Ala⁷⁷⁷ by Thr in the lodestar protein (Figures 3 and 5).

Figure 5.—

Domain organization of the Rad54, part of the LDS, and the A777T-LDS proteins. (A) The nucleotide triphosphate-binding so-called helicase motifs (I–VI) appear in shaded boxes and the E–N conserved domains are in open boxes in zebrafish Rad54A, a typical member of the Snf2 family of the helicase-related proteins (Flaus et al. 2006). (NLS, putative nuclear localization signal.) (B) The region including the J and the C boxes forms protrusion 2 that is composed of the α17 and α18 helices and the connecting short stretch of amino acids. Protrusion 2 was proposed to interact with the DNA (Thoma et al. 2005; Flaus et al. 2006). (See the inset and see www.sanger.ac.uk/cgi-bin/Pfam/swisspfamget.pl?name=P34739.) The presence of the B, the J, and the C boxes and the α17 and α18 helices is apparent in the LDS protein. The KK amino acids near the C box (labeled ** and also present in LDS) have been implemented in protein–DNA interaction (Thoma et al. 2005). In the LDS protein more amino acids compose the sequence that connects the α17 and the α18 helices as in Rad54. Presence of an α-helix is predicted inside this interconnecting region in the LDS protein. This α-helix became longer by two amino acids in the Horka^D encoded A777T-LDS protein as compared to LDS. The shaded scale at the bottom right illustrates the likelihood (as determined by the PSIPRED software: http://bioinf.cs.ucl.ac.uk/psipred/) that any amino acid is part of an α-helix.

Lodestar transgenes that carry the G²⁴²⁴ → A mutation (Figure 3) render females sterile. Although cuticle fragments appear inside 25–28% of their eggs, larvae never hatch (Table 1). Crosses between y v f mal females and +/+; TG^HD males yielded XX//X0 mosaics among the XX offspring (Table 1). Thus, Horka^D is a dominant lds mutant allele and the Horka^D phenotypes originated from the same mutation.

Phenotypic analysis of the horka^rvP2 null allele:

Cytological analysis revealed abnormal chromosome segregation in ∼30% (14/47) of the horka^rvP2/– egg primordia during the first meiotic division (Figure 1). Similarly, ∼38% (5/13) of the second meiotic divisions are unusual as shown by the abnormalities in both chromosome segregation and the formation of unusual spindles (Figure 1).

All eggs of the horka^rvP2/– females appear normal and are fertilized as in wild type, and although cleavage divisions commence inside ∼60% of the eggs, larvae never hatch. Once started, the cleavage divisions proceed more or less normally and cells may form over relatively large areas in the egg cortex and differentiate as indicated by the cuticle fragments that form inside 20.6% of the eggs (Table 1). Although the cuticle fragments are usually poorly differentiated, every larval cuticle landmark develops, albeit in different embryos.

Although daughter centrosomes separate appropriately, several of them lose the ability to nucleate microtubules. The centrosome defects lead to the formation of abnormal spindles, which then bring about a distorted arrangement of the chromosomes, a defect known as mitotic catastrophe (Figures 1 and 4; Sibon et al. 2000; Takada et al. 2003; Wichmann et al. 2006). While the nuclei close to the abnormal centrosomes drop from the egg cortex inside the egg cytoplasm, the centrosomes remain in place. Most of the free centrosomes nucleate microtubules and bring about further abnormalities by disturbing the nearby cleavage spindles. The impaired centrosome function may be related to one or more problems: DNA damage, incomplete replication of the DNA, abnormal chromatin condensation, and/or chromosome segregation. Thus, the loss-of-function mutant phenotype suggests involvement of lodestar in the maintenance of genomic integrity.

Figure 4.—

Impaired centrosome function develops in late cleavage embryos of the horka^rvP2/− females. Time-lapse optical sections were collected from embryos that derived from +/− (control) and from horka^rvP2/− females. The chromosomes were labeled by histone-RFP and appear in red, and the microtubules and the centrosomes were highlighted by Jupiter-GFP and are shown in cyan. Nuclei associated with abnormal centrosomes are within dashed circles. Note that while the nuclei drop into the interior of the embryo, the free centrosomes remain in the egg cortex. Bar, 10 μm.

Characteristic types of defects appear during spermatogenesis in the Horka^D/– males. As a consequence of nondisjunction, larger- and smaller-than-normal onion stage spermatid nuclei appear side by side (see supplemental Figure 2; cf. Szabad et al. 1995). Several of the sperm nuclei are displaced from their sperm tip position, and a good number of sperm tails bear no nucleus (see supplemental Figure 2).

Although the onion stage spermatid nuclei appear in the horka^rvP2/– males as in wild type, the sperm bundles are abnormal: individualization of the sperm is incomplete, a few of the sperm heads are dislocated, and the sperm head is missing from several sperm tails (see supplemental Figure 2). Yet some of the sperm must be functional as the horka^rvP2/– males are not completely sterile (supplemental Table 2).

The analysis of germ-line chimeras in Horka^D/+ and horka^rvP2/– flies:

The viability and sterility of the Horka^D/+ and the horka^rvP2/− females and reduced fertility of the males suggest that the function of lodestar is required only in the gonads. To determine whether function of the gene is required in the germ line or in the somatic components of the gonads, we constructed germ-line chimeras through the transplantation of pole cells. First, pole cells of Horka^D/+ embryos were transplanted into host embryos that did not have pole cells yet provided a normal environment for development and function of the donor pole cells (Table 2). Three of the female germ-line chimeras produced eggs, and the fate of embryos inside these eggs was essentially identical to that described for embryos of the Horka^D/+ females. Three sibling male germ-line chimeras were generated and then mated with y v f mal females. On average, 3.1% of their XX zygotes developed as XX//X0 mosaics (Table 2). Features of the chimeras clearly show that the Horka^D-induced defects originate from altered function of the germ-line cells.

We also used Horka^D/+ females and males as host for normal germ-line cells. Apparently fully functional germ cells developed from the transplanted pole cells in the Horka^D/+ environment and offspring derived from the chimeras that carried normal germ-line cells (besides their own) and Horka^D/+ soma (Table 2). Features of the latter types of germ-line chimeras not only revealed the germ-line autonomous nature effect of Horka^D but also showed that the Horka^D/+ gonadal soma functions normally.

In the second set of experiments, pole cells of horka^rvP2/− embryos were transplanted into Fs(1)K1237/+ host embryos. Of the developing chimeras three carried horka^rvP2/− germ-line cells (Table 3). They deposited normal-looking eggs from which larvae never hatched. Cuticle fragments were present in 21% of the eggs, as inside eggs of the horka^rvP2/− females (Table 1). We also transplanted normal pole cells into horka^rvP2/− host embryos and analyzed the developing female and male germ-line chimeras. The horka^rvP2/− flies produced offspring from the implanted germ-line cells (exclusively), showing that the horka^rvP2/− soma provides full support for the normal germ-line cells (Table 3). It appears that function of lodestar is primarily required in the germ line.

DISCUSSION

Nature of the Horka^D-encoded A777T-LDS protein:

Horka^D is an allele of lodestar, which encodes a member of the Snf2 family of the helicase-related proteins that are involved in transcription regulation, DNA repair, recombination, and chromatin unwinding (Flaus et al. 2006). The helicase motifs and the other conserved domains contribute to distinctive features in the Snf2 protein family (Figure 5). In Rad54, the only member of the family of known structure, two of the α-helices (α17 and α18) and a short interconnecting region compose protrusion 2, the part of the protein that interacts with DNA (Thoma et al. 2005; Flaus et al. 2006; Figure 5). The α17 and the α18 helices are present in the LDS protein but the interconnecting region is longer than in Rad54 and contains an α-helix (Figure 5). Horka^D is a G²⁴²⁴ → A transition that results in replacement of Ala⁷⁷⁷ by Thr in the interconnecting region. It appears that this amino acid replacement expands the α-helix by two amino acids (see supplemental Figure 1).

Possible function of the LDS protein:

The LDS protein is cytoplasmic during interphases of the cleavage mitoses, enters the nucleus during prometaphase, and becomes associated with the chromosomes throughout mitosis, suggesting an involvement of the LDS protein in chromatin/chromosome surveillance during mitosis (Girdham and Glover 1991; supplemental Figure 1). This idea is supported by the loss-of-function mutant phenotype in embryos of the horka^rvP2/− females: abnormal assembly of the chromosomes during meiosis and mitosis, formation of anastral centrosomes and abnormal spindle apparatus, failures in the cleavage mitoses, fallout of the abnormal cleavage nuclei, and eventual death of the embryos. Similar, if not identical, defects have been reported in embryos of those females defective in (i) spindle assembly checkpoint functions or (ii) the mitotic catastrophe avoidance mechanism (Castedo et al. 2004; Musacchio and Salmon 2007; Vakifahmetoglu et al. 2008; Yuen and Desai 2008). The latter mechanism operates through the activation of checkpoint kinase 2 (Chk2), whereby the damaged or the incompletely replicated DNA leads to Chk2 activation and resulting inactivation of the centrosomes and the spindles. These events in turn result in blocked chromosome segregation during anaphase and the eventual elimination of those nuclei from the embryonic precursor pool. The Chk2-based mechanism is especially important for maintaining genomic stability during genotoxic stress (Masrouha et al. 2003; Takada et al. 2003; Brodsky et al. 2004; Wichmann et al. 2006; Larocque et al. 2007). Defects in the Chk2-based mechanism cause mitotic catastrophe (Vakifahmetoglu et al. 2008).

LDS does not likely function in the spindle assembly checkpoint because—in contrast to the LDS and the Chk2 proteins—the spindle checkpoint proteins have been shown to bind to the kinetochores (Gillett et al. 2004; Musacchio and Salmon 2007). The abnormalities that emerge in embryos of the horka^rvP2/− females exhibit all the distinctive features of mitotic catastrophe. Largely identical defects have been described for checkpoint kinase 1 (grapes, grp), checkpoint kinase 2 (lok or maternal nuclear kinase, mnk), and Ataxia telangiectasia-related mei-41 mutant alleles (Masrouha et al. 2003; Brodsky et al. 2004; Royou et al. 2005; Takada et al. 2003, 2007; Wichmann et al. 2006; Larocque et al. 2007). Functions of the corresponding genes have been implicated in the G2/M checkpoint by “assaying” the status of the DNA and/or the chromatin and in the elimination of inappropriate nuclei from the pool that serves as a source of the blastoderm cells following the cleavage cycles (Takada et al. 2003; Larocque et al. 2007). It is possible that the LDS protein might be involved in the same pathway as Chk2, because a few of the embryos of the mnk/mnk; horka^rvP2/– females (lacking both the Chk2 and the LDS proteins) develop to adulthood (our unpublished results). Such an event never happens to embryos of the horka^rvP2/– females. The role of the LDS protein in chromatin surveillance and cell-cycle progression regulation, however, remains to be clarified.

The requirement of lodestar in the germ line and in the soma:

Remarkably, lds gene function is indispensable in the germ line but not in the soma: although flies develop normally without LDS protein, the meiotic divisions are abnormal in horka^rvP2/– females, as are the cleavage mitoses in their embryos (Girdham and Glover 1991 and this article). These defects are germ-line autonomous. In fact, complete or almost complete maternal-effect lethality is a characteristic feature of females that are homozygous for mutant alleles of the genes engaged in the G2/M transition control (Henderson 1999). For example, only ∼20% of embryos hatch from eggs of females that are both homozygous for the strong mnk mutant alleles and lack Chk2 function (Xu et al. 2001; Masrouha et al. 2003; Takada et al. 2003; Xu and Du 2003; Brodsky et al. 2004). Similarly, the grp homozygous females, which lack checkpoint kinase 1, are sterile; their embryos suffer from abnormal cortical nuclear divisions and do not cellularize (Yu et al. 2000; Jaklevic et al. 2006; Takada et al. 2007). Females homozygous for the Ataxia telangiectasia-related mei-41 strong mutant alleles are, in effect, sterile (Laurencon et al. 2003; Larocque et al. 2007).