hurku, a dominant mutation of drosophila, induces nondisjunction

Copyright 0 1995 by the Genetics Society of America

Hurku, A Dominant Mutation of Drosophila, Induces Nondisjunction and, Through Paternal Effect, Chromosome Loss and Genetic Mosaics

Jhos Szabad, * 9t Endre Mathe * and Jaakko Pur0

*Department of Biology, Albert Szent-Gyorgyi Medical University, H-6720 Szeged, Hungary, Institute of Genetics, Biological Research Center of the Hungarian Academy of Sciences, H-6726 Szeged, Hungary, and :Department of Biology, University of Turku,

SF-20500 Turku 50, Finland Manuscript received August 29, 1994

Accepted for publication November 17, 1994

ABSTRACT Fs(3) Horka (Horka) was described as a dominant female-sterile mutation of Drosophila melanogaster.

Genetic and cytological data show that Horka induces mostly equational nondisjunction during spermatogenesis but not chromosome loss and possesses a dominant paternal effect: the X, second, third and the fourth chromosomes, but not the Y , are rendered unstable while undergoing spermatogenesis and may be lost in the descending zygotes. The frequency of Horkeinduced chromosome loss is usually 2- 4% but varies with the genetic background and can be over 20%. The Xchromosome loss occurs during the gonomeric and the initial cleavage divisions. Loss of the X and fourth chromosomes shows no correlation. We propose, based on similarities in the mutant phenotypes with the chromosome destabilizing mutations nonclaret disjunctional and paternal loss, that the normal Horka+ product is required for function of the centromeres and/or nearby regions. Horka is a convenient tool for the generation of gynandromorphs, autosome mosaics and for the study of gene expression in mosaics.

R EPLICATION and transmission of the genetic ma- terial to daughter cells are achieved with remark-

able fidelity in all organisms. Although cell divisions have long been studied, rather little is known about the necessary proteins and how they control cell divisions (BAKER and HALL 1976). Genetic dissection proved to be an efficient approach in understanding the mechanism of cell division (ENDOW 1992, 1993; HAWLEY and THEURKAUF 1993). The cell division mutants of Dro- sophila have been useful in the elaboration of the cell cycle control mechanism (reviewed by GLOVER 1991 ) . Of particular interest are those Drosophila mutants that destabilize chromosomes. Chromosomes exposed to the mutant environment become unstable and may be lost in the descending embryos. The best known exam- ples of the chromosome-destabilizing mutants are nonclaret disjunctional, originally known as claret nondisjunctional ( cand) (WALD 1936; DAVIS 1969; SEQUIERA et al. 1989; YAMAMOTO et a l . 1989; ENDOW et al. 1990; MC- DONALD and GOLDSTEIN 1990; MCDONALD et al. 1990; THEURKAUF and HAWLEY 1992) , paternal loss ( p a l ) (BAKER 1975) and mitotic lossinducer ( mit) ( GELBART 1974; for reviews see HALL et al. 1976; ASHBURNER 1989 ) .

Horka is a new chromosome destabilizing mutation in Drosophila melanogaster. It was described as Fs(3) Hurka, a dominant female-sterile mutation that induces meiotic

Currespondingauthm:JAnos Szabad, Biology Department, Szent-Gybr- gyi Albert Medical University, H-6720 Szeged, Somogyi B. u. 4. Hun- gary. E-mail: [email protected] E-mail: [email protected]

Genetics 139: 1585-1599 (April, 1995)

defects as well as arrest during the initial cleavage divisions ( ERDELYI and S Z ~ A D 1989) . Hmka also induces mostly equational nondisjunction during spermatogenesis and possesses a dominant paternal effect: the chromosomes, except the Y, that undergo spermatogenesis in Hmka are rendered unstable and may be lost in the developing zygotes. Hm-ka is a gain-of-function mutation and the mutant phenotypes are caused by the mutant gene product ( ERDELYI and SZABAD 1989). The Horka- associated dominant mutant phenotypes stem from one mutation because they concomitantly disappear in all the eight revertants of Hmka. This paper provides genetic, developmental and cytological characterization of the dominant paternal effect of Hurka.

Genetic mosaics generated by chromosome elimination have been useful tools in cell lineage studies and in analysis of gene expression during ontogenesis. Mu- tations that destabilize chromosomes ( can", pal and mit) and the unstable ring-Xchromosomes, in particular I n ( 1 ) w'", are of special importance in the generation of genetic mosaics (for reviews see HALL et al. 1976;

JANNINC 1978; ASHBURNER 1989). Horka, through its chromosome destabilizing effect, proved to be a useful tool for the generation of gynandromorphs (SZABAD and NOTHIGER 1992) and autosome mosaics.

MATERIALS AND METHODS

Fs(3)Horka (Horka) is a dominant female-sterile mutation of D. melanogaster ( E m E m and SZABAD 1989). Horka is a gain- of-function mutation because it can be reverted to loss-of- function alleles (EmELn and SZABAD 1989) . The Horka mu-

1586 J. Szabad, E. M5th6 and J. Pur0

tant phenotypes originate from one mutation, because they simutaneously disappear in the revertants of Horka. The Hmka locus resides in the cytological region 87B (J. S z m m et al., unpublished data). Horka is not alleleic with the other chromosome destabilizing mutations. The Horka-carrying third chromosome is labeled with the recessive marker mutations mwh and e . Horka is kept in a self-propagating stock in which T(l;?) OR60/ TM?, Sb Srr females are mated with Horka/ TM?, Sb Srr males. The T(l;?I) OR60 rearrangement induces dominant male lethality. For an explanation of the genetic symbols, see LINDSLEY and ZIMM (1992).

Generation and detection of mosaics: Horka/ TM?, Sb Srr males taken from the above stock (or following outcrosses) were mated with females that carried X-linked recessive marker mutations. Mosaics were detected in the Horkaderived offspring by identification of gynandromorphs (female// male), as well as haplo-4 mosaics (JANNING 1978; ASHBURNER 1989). To detect the loss of major autosomes, Horka males were constructed that carried a CyOp or a TM2O chromosome with a transgene. The transgene ensured an "all over" expression of 0-galactosidase ( BELLEN et al. 1989). The transgene in the CyOp chromosome is already expressed at the cellular blastoderm stage, whereas the TM2O-linked transgene is expressed from the stage of germ band elongation. The diplo//haplo autosome mosaics were detected after fixing and staining for 0-galactosidase the embryos descending from the above males and y 71 f mal females.

For the detection of Ychromosome loss, Hcrrka males canying the y+Y or the w + Y chromosome were mated with X'/X' or X7"/ X"' females homozygous for the X-linked recessive marker mutations y or 7u, respectively. [The y+ Y and the 7u+ Y chromosomes carry the y + and the w+ segment of the Xchromosome ( LINDSLEY and ZIMM 1992) .I Loss of the g+Y or the 7u+Y chromosome should lead to the formation of XJy+Y//XJO or Xr"7u+Y// X " ' 0 ~ a l e mosaics.

Loss of the XY chromosome was studied i n 2 cross between X"'/ X"' females aRd males that carried the XY chromosome and Hmka. The XY chromosome is XYLYS, yz S U ( W " ) 7u". It is an intact Y chromosome with all the X chromosope euchro- matin ( LINDSI.EY and ZIMM 1992.) Loss ,Of the XY chromosome should lead to the formation of X'"XY//X"'O gynandromorphs, many of which are expected to possess w + / / w mosaicism.

Detection of nondisjunction: To test whether Hurkn can induce chromosome loss and/or nondisjunction during spermatogenesis, we made use of the compound second chromosome C(2)EN, bw sp ( GONZAIXZ et al. 1989). In this system, cn h u / ~ r l dp c l ; Hurka/ TM? (or + ) males were mated with C(2)EN, bw s p females. Oocytes of the C(2)EN, bzu sp females carry either two or none of the second chromosomes. Viable offspring do not result when such gametes are fertilized by regular sperm. However, viable offspring develop when the diplo-2 and nullo-2 oocytes fuse with aneuploid nullo-2 or diplo-2 sperm, respectively. The phenotype of the descending progeny allows conclusions to be made about the mechanism generating the aneuploid sperm.

Cytology: The chromosomes, the spindles and the sperm tail were made visible in the testes and in the newly fertilized eggs by double staining with Feulgen and Giemsa ( PURO and NoKKAIA 1977; P~JRO 1991). The opened male abdomens were immersed into a hypotonic (0.075 M ) KC1 solution for 10-13 min and fixed in a mixture of ethano1:chloroform:acetic acid 6:3:1 for 5 min. The abdomens were subsequently transferred to 99% ethanol and carried through reduced con- centrations of alcohol to distilled water. A 30-min soaking in 1 N HCl followed first at room temperature and then at 60" for 8 min to achieve hydrolysis. Feulgen staining was carried

out by treatment with diluted Schiff's reagent for 2-4 min. The abdomens were then transferred to a drop of distilled water, the testes were dissected and transferred to a slide. The water was replaced by 45% acetic acid. The testes were covered with a coverslip and gently squashed. The coverslip was removed on dry ice. The preparation was first dehydrated in 99% ethanol and then immersed in 100% methanol for 5 min. The preparations were dried and stained in 2% Giemsa in phosphate buffer (pH 6.8) for 15-30 min. Permanent preparations were made after a rinse in distilled water, drying and mounting in Entellan.

Preparations of fertilized eggs were made using a modifica- tion of the above procedure. Newly deposited eggs and those squeezed out from the uterus were put immediately into fresh fixative for 15-30 min. Ethanol (95% or 99%) was added to double the volume. Fixation was allowed to proceed for another 30 min. The rest of the procedure was as described above. Eggs were cut into an anterior and a posterior piece in distilled water and the chorion was removed. After dehydra- tion in ethanol, the slides were immersed into glacial acetic acid for 20 sec. Staining was carried out in 4% Giemsa for 30 min.

Analysis of gynandromorphs: The adult gynandromorphs were recovered under a dissecting microscope based on the morphology of the sexually dimorphic structures and/ or un- covering the X-linked recessive marker mutations in the X 0 male tissues. A random sample of 80 Hwkuderived gynandromorphs were mounted and analyzed under the compound microscope ( SZABAD 1978). For comparison, 80 "traditional" gynandromorphs were generated by the unstable ring-Xchro- mosome Zn(I) w"','. They were also mounted and analyzed. The haplo-4 mosaics were detected on the basis of the thin short bristles and the rough eye phenotype (ASHBURNER 1989). Several Horka-derived gynandromorphs were also identified, using standard immunological methods, at the blastoderm stage by the Sex-lethal antibody that identifies female nuclei ( BOPP et al. 1991 ) . The blastoderm gynandromorphs were mounted and analyzed to determine the relative areas covered by the female and male cells.

RESULTS

The dominant paternal-effect of Horka: loss of the paternally derived X chromosome leads to formation of gynandromorph: A cross between y u f mal females and Hwka males yielded 20.5% gynandromorphs and 2.5% haplo-4 mosaics among the XXzygotes and initi- ated analysis of the Horka-related mosaicism. The following observations show that gynandromorph formation is indeed the consequence of the Hwka mutation. Gynandromorphs did not develop among the progeny of y u f mal females and three different types of males: ( 1 ) mwh e / TM?, ( 2) Fs(?) / TM3 or ( 3) revertants of Horka (Table 1 ) . [ Horka, like the Fs(3) mutations labor^'^', Karta12'", T011'~" and Tomaj16d, was induced on an mwh e-labeled chromosome ( ERDELYI and SZABAD 1989) .]

There were 1035 gynandromorphs recovered during the course of the present study, that originated from X J X zygotes. ( X J and X stand for the y-labeled maternally and the y + allelecarrying paternally derived X chromosomes, respectively). The male parts were in- variably XYO, indicating that the X chromosome insta-

Paternal Effect and Chromosome Loss 1587

TABLE 1

Features of Horkainduced gynandromorph formation

Zygotes Parents

Frequency of No. of Females Males xx xx//xo xx//xo= (%) experiments

Any X chromosome Horka/any third chromosome 9365 262 1.7 -+ 1.8 11 Any X chromosome Horka/ TM3, Sb SeP 1278 41 3.2 ? 2.8 4 y v f m a l Horka/any third chromosome 4010 152 3.8 2 1.5 12 y v f m a l Hmka/TM3, Sb SeP 1614 335 20.7 ? 4.8 11 y v f m a l Hmka/+ or TM3, Sb SeP 4304 432 10.0 ? 3.6 17 Controls

y v f m a l mwh e/ TM3, Sb SeP 1309 0 - 5 y v f m a l Fs (3)/TM3, Sb SePb 1129 0 - 4

y v f m a l Horkaw/ TM3, Sb Ser 3301 1 0.03 9

Nine different X chromosomes were used besides the one labeled with y v f mal ( y w sn', y w", w, fs(1)KlO w, "5, y w CLI v f car, fs(1)KlO w f mal, wild type and one with a lacZ transgene). Thirty different third chromosomes were used, including 11 nonlabeled ones, four balancer chromosomes ( TM3, TM2, TM6B and CXD from different strains) and three with lacZ transgenes. The TMS, Sb SeP chromosome derives from the Fs (3)/T(l; 2)0R60/TM3, Sb Ser stock. Hmka" includes one EMS and eight P- element-induced revertants. Pooled data.

"Values are means ? SD. Pooled data from the dominant female-sterile (Fs) mutations labor^'^", Karta121g, Toll18a and Tomaj'6d that were induced on

mwh e-labeled chromosomes (ERD~LYI and SZBAD 1989).

bility is an inherent nature of the Hm-ka-derived chromosomes.

In Drosophila eggs, the first mitotic (gonomeric) division occurs in parallel in both haploid genomes after fertilization. Chromosomes of maternal and paternal origin mix at the spindle poles and form two zygotic nuclei ( LIN and WOLFNER 1991 ) . The zygotic nuclei undergo a series of 13 rounds of rapid synchronous cleavage divisions without cytokynesis, before the formation of the cellular blastoderm ( GLOVER 1991 ) . The existence of the y male parts clearly shows that the Hm-kct-derived Xchromosome was either lost or did not segregate to the nuclei populating the male cells of the gynandromorphs.

Gynandromorphs may originate through either of two mechanisms: loss of an Xchromosomes during the gonomeric or the cleavage divisions, which leads to the formation of XYX//XYO type of gynandromorphs, and mitotic nondisjunction may result in the formation of XyXX//XYO gynandromorphs. The following para- graphs present evidence that the Hm-kninduced gynandromorphs originate through loss and not nondisjunction of the paternally derived X chromosomes.

It is known, from analysis of XXX// XX mosaics, that the first and second legs as well as the wings are very seldom composed of XXXcells due to the XXX-associated lethality mapping to the ventral thoracic region ( SCHCJPBACH et al. 1978) . Viability of the thoracic structures is not reduced in the Horkninduced gynandromorphs showing that they are XYX/"XYO and not XyXX//XyO type and hence originated after the loss of the Horkaderived X chromosome (Table 2 ) .

The XYXX//XyO zygotes are expected to live at least to the end of the larval life, because the XXX diploid metafemales die in the late larval and pupal stages ( LINDSLEY and ZIMM 1992) . There were 12 late third instar gynandromorph larvae isolated, showing y + / / y mosaicism in the head skeleton, from the cross between y ZI f mal females and Hm-ka/ TM3, Sb Ser males. They were likely to have had mosaic central nervous systems and/ or imaginal discs ( c j JANNING 1978) . Cytological analysis clearly showed that their brains and/ or imaginal discs were composed of only XX and/or X 0 and not of XXX cells. Hence, the Horkninduced gynandromorphs are indeed XyX//XYO types and originate through elimination of the Horkaderived X chromosome.

To visualize cytological loss of the chromosomes, eggs of the y ZI f mal females that had been mated with Hwka males were fixed within a few minutes after fertilization. The eggs were double stained with Feulgen and Giemsa and analyzed for the detection of lost chromosomes. As shown in Figure 1, chromosomes were identified lagging between the telophase chromosomes in the for- mer equatorial plane in both the gonomeric and the subsequent cleavage divisions. It is generally accepted that the lagging chromosomes are lost during cleavage divisions ( c j WALD 1936; ZALOKAR et al. 1980). The presence of the lagging chromosomes indicates that the mosaics originate through chromosome loss and not nondisjunction. We did not detect any lagging chromosome in the control experiment where y u f malfemales were mated with mwh e / TM? males.

The paternal effect of the Horku-induced gynandro-

1588 J. Szabad, E. Mathi. and J. Pur0

TABLE 2 Frequencies (%) of adult structures in XXX//XX, Zn(l)uYc and Horkiiduced mosaics

Structure

Genetic Tergites Sternites Type of mosaic constitution First leg Second leg Wings (T2-T6) (Sl-S5)

"x// xx xxx 2 3 8 33 35 xx 70 68 67 47 55 =// xx 28 28 26 20 10

In(l)wK-induced gynandromorphs xx 45 44 46 51 59 x0 30 31 34 34 33 xx// x0 25 25 20 15 8

Hmkninduced gynandromorphs xx 43 47 42 50 56 x0 35 37 45 40 39 xx//xo 22 16 13 10 5

Data related to XXX//XX mosaicsim is from SCHUPBACH et al. 1978.

morph formation, an indication of Xchromosome loss, is nicely shown by the following observation. Among the progeny of y u f rnal females and Hwka/ TM3, Sb Ser males, the frequencies of Horka/ + and TM?/ + gynandromorphs were not significantly different (46,' 173 us. 29/ 150; P > 0.05) ( JSASTENBAUM and BOWMAN 1970).

Unlike the case of In (1) wyc, there is no dominant lethality associated with the Hwka mutation. Males and females developed with -equal frequencies among the progeny of y u f rnal females and Horka/ TM?, Sb Ser males. The ma1e:female ratio was 0.97 -+ 0.10 (mean ? SD ) in those 10 crosses where the number of males was recorded. When, however, the y u f mal females are mated with In (I) wU'"/yf Y males, an excess of males develop (5816 males and 2493 females) ( SZABAD et al. 1979) due to the dominant lethality associated with the In(1)w'" chromosome (for a review see ASHBURNER

1989 ) . Parameters of the X chromosome loss: We analyzed

55 crosses during the course of this study (Table 1 ) . When males carrying Hwka and any of the eight different third chromosomes were mated with females that carried eight different types of X chromosomes, 1.7% of the zygotes were gynandromorphs.' The frequency of gynandromorphs was 3.2% when the X / Y ; Hwka/ TM?, Sb Ser males were taken from the Horka/ T(I;?) OR60/ TM?, Sb Serstock and the females carried one of four different types of X chromosomes. The frequency of gynandromorphs was 3.8% in those 12 crosses in which the females were y u f mal homozygous and the males carried Hwka and one of nine different chromosomes, including TMZ, TM6B and TM?, Sb Ser.

' The X chromosome did or did not carry recessive marker mutations in the aforementioned experiments. The variation in the frequency of gynandromorphs can be accounted for to some extent by the fact that when the X chromosome is not labeled with recessive marker mutations only 83% of the gynandromorphs can be identified: those that show mosaicism in the sexually dimorphic structures.

These males originated from outcrosses with different types of females and Horka males. The frequency of gynandromorphs exceeded 20% when y u f mal females were mated with X/Y; Horka/ TM3, Sb Ser males, taken from the Hmka/ T(l;?) OR60/ TM3, Sb Ser stock. The above experiment was repeated 11 times over a span of 5 years. This last type of cross was also carried out at 18 and 29" and yielded 10.3 (11/107) and 6.7% (14/ 209) gynandromorphs, respectively. These frequencies are significantly different from the 20.7% frequency (335/1614; Table 1; P < 0.05 and P < 0.01, respectively; chi-square test) determined at 25" and indicates the temperature-sensitive nature of the Hwka-induced X-chromosome loss.

The frequency of gynandromorphs was -10% in those 17 experiments where the y u f mal females were mated with Hwka males that originated from outcrosses between Horku-carrying males taken from the Hwka/ T(I;?) OR60/ TM?, Sb Serstock and five different types of females or carried, besides Horka, the TM3, Sb Ser chromosome that derived from the T(1;3) OR60 stock after three to four rounds of outcrosses. Results summa- rized in Table 1 clearly show that Horka brings about X chromosome instability that leads to the formation of X X / / X O mosaics and refer to at least two factors necessary for high frequency of chromosome loss, one being present in the Horka/ T(1;3 ) OR60/ TM?, Sb Ser and the other in the y u f mal stock.

The frequency of gynandromorphs dropped to 11.5% (59 / 514) in crosses where the females were y u f mal; Cy/ Pm; CxD/ TM? and the Horka males derived from the Horka/ T ( 1;3) OR60/ TM3, Sb Ser stock. Simi- larly the frequency of gynandromorphs dropped to -10% when the y u f mal females were mated with X / Y; Cy/Pm; Horka/ TM3, Sb Sermales. When y w cu u f car or y w f mal (each sharing three marker mutations with the y u f rnal chromosome) females were mated with Horka males from the Horka/ T(I;?) OR60/ TM?,

Paternal Effect and Chromosome Loss ~ ~~

1589

FIGURE 1.-Cytology of early embryogenesis in eggs of y u f mal females that had been fertilized by Horlza sperm. Gonomeric division, at prometa (A) and anaphase (D) , showing side by side the maternally and the paternally derived chromosome sets. Please notice the monopolar nature of the spindle apparatus during the gonomeric division. The A1 inset is an enlargement of the gonomeric spindle apparatus. (B) Second embryonic division in anaphase. The B1 inset Rshows an enlargement of the right- hand spindle. (C) Gonomeric division at prometaphase. (E-G) Telophases in the cleavage divisions. The arrowheads point to the centrosomes. Note that one of the centrosomes is attached to the sperm tail (st) in A and B ( CJ KARR 1991 ) . Arrows point to displaced or lagging chromosomes prone to loss. Scale bar: 20 pm on A and B and 10 pm on Al, B1 and C-G.

Sb Serstock, 15.5% (32/207) and 20.3% (35/172) of the XX zygotes were gynandromorphs. Attempts failed to identify the factor ( s ) of the y u f mal chromosome responsible for the high frequency of chromosome loss. Recombinant chromosomes between a wild-type and the y u f mal chromosome did not bring about high frequency of chromosome loss. A similar experience was reported following mapping factors that modify the frequency of pal-induced chromosome loss (BAKER 1975). The above data demonstrate that unknown genetic factors contributed by both parents can alter the frequency of Xchromosome loss in the zygote.

The Drosophila females store sperm for -10-12 days after copulation. When progeny derived from the stored sperm of y v f mal females and Hmka males, the frequency of gynandromorphs did not change significantly during a 10day sperm storage period ( P < 0.05; Table 3 ) . This result shows the long-lasting action of the mutant Hmka gene product. The haplo-4 mosaicism: There were 269 haplo-4 mo-

saics and 781 gynandromorphs recovered in those 39

TABLE 3

The relationship between sperm storage and the frequency of gynandromorphs

Zygotes

Days after Frequency of copulation Total X X X X / / X O X X / / X O " (%)

0-2 820 168 20.5 2 2.1 2-4 56 1 138 24.6 2 1.9 4-6 397 105 26.4 2 4.2 6-8 189 44 8-1 0

23.3 t 3.3 67 15 22.4 2 3.5

Total 2034 470 23.1 2 1.6

y u f mal females were mated with Horka/TM3, SI, SPT males that derived from the Hmlm/T( 1 ; 3)OR60/ TM3, S I 1 SPT stock. Three parallel samples were studied. The males were dis- carded after 2 days, and eggs were collected subsequently from the females in 2-day intervals.

"Average of three samples 2 SD.

1590 J. Szabad, E. Math6 and J. Puro

TABLE 4

Features of Hmka;induced haplo-4 mosaic formation

Zygotes

Parents

Females Males Total X X

y v f m a l Controln 5740 y v f m a l Horka/ TM3, Sb S e P 735 Any X chromosome Horka/TM3, Sb SeP 2007 yvfmal Horka/any third chromosome A 3678

B 1690 C 1120

Any X chromosome Horka/any third chromosome 4741

Haplo-4 mosaics

Total %

0 16 2.1 ? 0.4 25 1.4 ? 0.9 61 1.6 ? 0.7 62 3.8 ? 0.7 78 6.9 ? 0.7 27 0.7 ? 0.5

-

No. of experiments

4 7 9 6 7 6

~ ~~ ~ ~ ~~ ~~ ~~

The TM3, Sb S e P chromosome derives from the Es(3)/T( 1; 3)OR60/TM3, Sb Ser stock. A, at least third generation outcross;

Pooled data from mwh e/TM3, Sb Ser, the dominant female-sterile mutations labor^'^', Karta121g, Toll1'" and Tomaj16" as well B, second generation outcross; C, first generation outcross.

as the nine Hwka revertants.

experiments where haplo-4 mosaics were scored.' (Only the XX, haplo-4 mosaics were recorded. The XY; haplo-4 mosaics develop with similar frequencies as the X X haplo-4 ones.) Formation of haplo-4 mosaics is a consequence of the Horka mutation: haplo-4 mosaics did not develop among the progeny of (1) males carrying the mwh &labeled chromosome, ( 2 ) either of the four Fs (3) mutations listed in Table 1 and ( 3 ) any of the Horka revertants. The frequency of the haplo-4 mosaicism varied between 0.7% and 2.1% among the Horkaderived XXzygotes in 26 of the 39 crosses (Table 4) and did not seem to depend on the male and female genotype. However, when the Hmka males carried the "sensitive" fourth chromosomes the frequency of haplo-4 mosaics increased significantly ( P < 0.01; Table 4 ) . The sensitive fourth chromosomes derived from eight different stocks. The frequency of haplo-4 mosaics was 6.9% and 3% when the Hm-ka males were derived from the second or first generation outcrosses, respectively (Table 4, B and C ) .

The relationship between elimination of the X and the fourth chromosomes is linear and is shown on Fig- ure 2. In six crosses the frequency of haplo-4 mosaics exceeded those of the gynandromorphs. The Horka males were derived from first generation outcrossings in six of the eight crosses and mothers of the Horka males provided the sensitive fourth chromosome. (The gynandromorph/ haplo-4 mosaic ratio was 0.6 2 0.2; r = 0.90; P -=% 0.01; Figure 2, line A ) . In 17 crosses the

'Although detection of gynandromorphs is based on coexistence of female and male structures over the entire body, the haplo-4 mosaics were identified based on short head and notum bristles and/or the rough eye phenotype (ASHBURNER 1989). Haplo-4 mosaicism restricted to the abdomen went undetected. Of the 80 Hmka-induced gynandromorphs analyzed in a compound microscope, 55 (71%) carried male clones on the head and/or the notum, leaving the abdomen composed of entirely male or female structures. This observation suggests that -29% of the haplo-4 mosaics went undetected.

frequency of gynandromorphs slightly exceeded those of the haplo-4 mosaics. The Hm-kmarrying males were derived from the second and / or third generation after outcrosses between Horka males and the sensitive fourth chromosome-providing females. (The gynandromorph / haplo-4 mosaic ratio was 2.6 % 0.6; r = 0.92, P G 0.01; Figure 2, line B ) . In 14 crosses many more gynandromorphs developed than haplo-4 mosaics. In 7 of the 14 crosses the Hm-ka males were directly taken from the

6"i 4" -

-7"

I I I

I I

1"

I

- r I

I I I 5 10 15 20

% Gynandromorphs

FIGURE 2.-The linear relationship between the frequency of gynandromorphs and the frequency of haplo-4 mosaics. The gynandromorph/haplo-4 mosaic ratio was 0.6 ? 0.2 in 6 ( 0 , line A ) , 2.6 ? 0.6 in 17 (m, line B) and 7.7 ? 2.5 in 14 of the crosses (A, line C) .


Horka/ T( 1;3) OK60/ TM3, Sb Serstock and mated with y v f mal females. In the remaining seven crosses the males were derived from outcrossings with different females that did not carry the sensitive fourth chromosome. (The gynandromorph/h-4 ratio was 7.7 ? 2.5; r = 0.88, P 0.01; Figure 2, line C ) . The slope of line B is the average of lines A and C (Figure 2) .

The above results show the presence of factor ( s ) in the different genotypes that render the fourth chromosomes highly or only moderately sensitive to Horka action. The linear relationship between the Horka-related X and fourth chromosome instability points to a common source for loss of the X and the fourth chromosomes.

Independent loss of the X and the fourth chromosomes: Unlike in cand and pal, where loss of the Xand fourth chromosomes mostly takes place simultaneously and, as a rule, the male parts of the gynandromorphs are also haplo-4 mosaics, the Horka-induced X and fourth chromosome losses are independent. In one experiment a total of 1695 zygotes were XXat fertilization. Of the 1695 zygotes, 130 (7.7%) were gynandromorphs. Loss of the fourth chromosome took place in 55 (3.2% ) of the zygotes. It was expected, based on frequencies of X and fourth chromosome losses, that four of the zygotes would bear X 0 and haplo-4 tissues simultaneously. X 0 and haplo-4 tissues were simultaneously present on 3 of the 1695 zygotes. Borders of the X 0 and haplo-4 clones did not coincide in two of the above three mosaics. In one case, part of the X 0 clone was also haplo-4, indicating that loss of the fourth chromosome took place in a nucleus that had already lost the X chromosome.

Loss of the major autosomes: Loss of the second chromosome was studied in embryos descending from y v f mal females and males that carried a nonmarked Xchromosome and a CyOp chromosome with a /3-galactosidase transgene ensuring "all over" expression of the enzyme at the cellular blastoderm stage. In this cross, 50% of the embryos should stain blue. However, if Horka induces loss of the paternally derived CyOa chromosome, a few of the zygotes will be diplo-2// haplo-2 mosaics and show blue //white mosaicism. Seven (3.4%) of the 216 embryos with the transgene showed blue //white mosaicism (Table 5 and Figure 5 ) . The existence of blue // white mosaics clearly shows that in addition to the Xand the fourth chromosomes, Horka also renders the second chromosomes unstable. (The blue //white mosaic pattern did not develop in the 603 blue-staining embryos that descended from the control males who carried CyOp but not Horka.) A proportion of the sibling embryos developed to adulthood. The frequencies of X and fourth chromosome loss was 3.4% and 7.8%, respectively (Table 5 ) .

The Horka males carried the TM2O chromosome with the P-galactosidase transgene in experiments where loss

of the third chromosome was assayed. Two (0.8%) of the 251 TM2O-carrying embryos possessed blue // white mosaicism, indicating that Horka also renders the third chromosome unstable (Table 5 ) . (None of the 408 TM2Ocarrying control embryos showed blue// white mosaicism.)

Detection of the diplo-3// haplo-3 mosaicism was limited. Expression of the ,&galactosidase transgene commences only during germ band elongation, at a stage when viability of the haplo-3 cells is already reduced. A comparison between the frequencies of X, fourth, second and third chromosome loss is compli- cated by the low viability of the haplo-2 or haplo-3 cells. Monosomy for either chromosome 2 or 3 leads to embryonic lethality ( WRIGHT 1970). The stage of death of the diplo-2// haplo-2 and diplo-3// haplo-3 mosaics may also depend on the size and/or position of the haplo clones.

Horlu-derived Y chromosomes are not lost: Six experiments were conducted to evaluate whether Horka induces loss of a y+Ychromosome (Table 6 ) . First, Xy/ Xy females were mated with X/y+Y Horka males. The body of the regular XY/yCYmale offspring is wild type. Loss of the y + Y chromosome in the Xy/y+Y zygotes should lead to the formation of XYy + Y// XyO wild- type //yellow male mosaics. None of the 3267 Xy/y' Y males were wild-type// yellow mosaic (Table 6 ) .

The Horkaderived Xchromosome was nonmarked in the above experiments. Its loss could be detected in the sibling XyX zygotes. Of the 5557 XyX zygotes 237 (4.2%) were gynandromorphs and 99 (1.8%) haplo-4 mosaics. Assuming similar frequencies for the elimination of the X, fourth and the y + Y chromosomes, formation of - 139 or 58 wild-type // yellow mosaic males was expected. As stated above, none was detected. There were 68 yellow males identified in the above cross besides the 3267 wild-type looking ones. They all were sterile, most likely XyO males.

In a second experiment, loss of a w+ Y chromosome was studied: X w / X" females were mated with X/ w+Y; Horka males (Table 6 ) . Of the 858 X7"/ wf Y males, none was X7"w+Y//Xw0 mosaic showing red//white eye mosaicism. There were 1014 X"X sibling zygotes recovered. Forty-five (4.4%) and 12 ( 1.1% ) were gynandromorphs and haplo-4 mosaics, respectively. Based on these frequencies, formation of 38 or 9 red//white male mosaics was expected. Again, none developed. The above results show that Horka-induced loss of the Y chromosome is at least an order of magnitude less than that of the X or any autosome and may in fact never occur. ~

Loss of the XY chromosome: Lack of Y chromosome loss could be the consequence of absence of factors that render the Ychromosomes susceptible to Horka action or Ylinked factors that may inhibit the Hmka destabilizing action. These alternatives were investigated by studying

1592 J. Szabad, E. Mhthe and J. Puro

TABLE 5

Features of Hothindwed major autosome loss

Diplo//haplo XX zygotes Diplo

Male autosome mosaics

Haplc-4 Chromosome type zygotes" Total % Total xx/ x0 % mosaic %

Second Control 603 0 - Not determined

Third Control 408 0 - Not determined Hmka 216 7 3.4 204 7 3.4 16 7.8

Hmka 251 2 0.8 256 8 3.1 18 7.0

y uf mal females were mated with males that carried a &galactosidase transgene on the second or on the third chromosome

a Only the transgene-carrying zygotes were included. (see MATERIALS AND METHODS).

the action of H d a on an XYchromosomz. In this experiment, X"/ x," females were mated with XY; Horka males. Loss of +e XYchromosome should lead to the formation of X"XY//X"O gynandromorphs with white (mal%) clones in the red (female) eye. Among the 989 X"XY zygotes, 21 (2.1%) were gynandromorphs, as identified based 02 the sexually dimorphic structures. Nine of the 21 X"XY//X"O gynandromorphs shcwed red//white eye mosaicism. Formation of the X"XY//X"O mosaics clearly indicates that loss of the Ychromosome takes place when it is attached to the X chromosome. We thus con- clude that the X chromosomes and autosomes, but not the Y chromosome, cany factor ( s ) that render them susceptible to Horka action.

The origin of the X 0 males: The Y chromosomes carried the y + or the w+ segment of the Xchromosome in 11 crosses. The partner females were homozygous for the marker mutations y or w . There were y or w exceptional males in addition to the regular Xy/y+Y or X"/ w+Yones. All the exceptional males were sterile, thus likely XO. As shown in Figure 3, the frequency of X 0 males is directly proportional to the frequency of gynandromorphs ( r = 0.96, P < 0.01 ) . This result indicates that the formation of X 0 males and gynandro-

A

morphs has a Hwknrelated common source. X 0 males also develop among the progeny of males carrying the unstable ring-X chromosome Zn(1)~''~''. In a cross where y v f mal females were mated with In (1) wl'"/ y + Y males, 712 (10.9%) of the 6528 males were X 0 ( SZABAD et al. 1979) . Of the sibling 2493 Xy/ Zn (1) w"" zygotes, 854 (34.3%) were gynandromorphs. The Zn(1)wuc-related X 0 male and gynandromorph frequency data fit the Horka-related ones (Figure 3) and suggest a common route for the formation of X 0 males in the two systems.

Because it is very unlikely that the X 0 males originated through the simultaneous loss of the unstable X chromosomes from both zygotic nuclei during the gonomeric division, it may be suggested that the X 0 males developed due to fertilization of the X-bearing eggs by nullo-X sperm.

Horka induces mostly equational nondisjunction during spermatogenesis: The Horknderived X 0 males apparently originated through the fertilization of the X- bearing eggs by nullo-Xsperm. The nullo-Xsperm may originate due to chromosome loss and/or nondisjunction during spermatogenesis. To decide between the above possibilities, X/Y; cn bw/ ed d p 61; Hwka/ TM3,

TABLE 6

The lack of Hurkminduced Y chromosome loss

Offspring

Males Haplo-4

Parents Mosaic Gynandromorphs mosaics

Females Males Regular Obs. Expb X 0 Females Total %" Total %"

x/-Y X/y+K Hmka 3267 0 139; 58 68 5221 237 4.2 99 1.8 r/r X/w'Y; Hmka 858 0 38; 9 20 957 45 4.4 12 1.1

Six crosses, with females and males of different genotypes, were studied to assay possible loss of the y+Y chromosome. The maternally derived chromosomes were labeled with y, a recessive marker mutation. Possible loss of the w+Y chromosome was analyzed in two crosses. The maternal chromosomes were w labeled.

% of X X zygotes. "The expected figures were calculated based on the frequency of gynandromorphs and haplo-4 mosaics, respectively.


-1 0 1

log % Gynandromorphs

FIGURE 3.-The linear relationship between the frequency of gynandromorphs and X 0 males ( r = 0.96, P < 0.01 ) . Notice that the value related to the unstable ring-Xchromo- some Zn(1) wUc (El in the top right corner) fit nicely the line based on the Horka data ( 0 ) .

Sb Ser (or + ) males were mated with C(2)EN, bw sp females. Progeny derive from the above cross only if the males produce nullo-2 and/or diplo-2 sperm. bw sp flies originate after fertilization of diplo-2 oocytes by nullo-2 sperm. Wild-type, cn bw and ed dp cl flies originate after fertilization of nullo-2 oocytes by diplo-2 sperm. The wild-type class is an indicator of reductional nondisjunction. The cn bw and the ed dp cl classes originate from equational nondisjunction and should theo- retically be present in equal numbers.

A single offspring developed in the control (Table 7 ) . The 53 Hwk&carrying males yielded altogether 1085 offspring, showing that Hwka males do indeed produce sperm with unusual numbers of chromosomes. Only 35 wild-type flies developed among the 1085 off-

spring (Table 7 ) . There were also 382 cn bw and 122 ed dp cl offspring flies recovered, indicating that Hwka induces mostly equational nondisjunction. (The difference between the frequencies of the cn bw and the ed dp c l flies presumably stems from the difference in their viabilities.)

The ratio of nullo-2 and diplo-2 sperm is 1:l after nondisjunction during spermatogenesis. If the nulle2 sperm originate only from nondisjunction, the number of the bw sp progeny flies, indicators of the nullo-2 sperm, should be similar to the sum of the wild-type, the cn bw and the ed dp cl classes. If, however, Hwka induces loss of the second chromosomes during spermatogenesis, an excess of the bw sp progeny flies should develop. Because the number of bw sp flies (546) was similar to the sum of the others (539) and no excess in the bw sp class emerged (Table 7 ) , we propose that Hwka does not induce detectable frequency of chromosome loss during spermatogenesis.

The nullo-2 sperm are likely to be the result of nondisjunction and not chromosome loss. This assumption is supported by the following observation. Both products (the nullo-2 and the diplo-2 sperm) of the equational nondisjunction were recovered from every one of the 53 tested males (Table 7) : all 53 yielded bw sp offspring and in addition either cn bw (14 males) or ed dp cl ( 2 males) or both cn bw and ed dp cl flies (37 males). In the case of loss of the second chromosome, some of the males should have given rise to only bw sp offspring.

Cytological analyses confirmed the presence of aneuploid sperm in the testes of Hwka males. The “onion” stage spermatids were analyzed first on squashes of the Hwka males taken from the H d a / T( 1;3) OR60/ TM3, Sb Ser stock. The volume of the onion stage spermatids is directly proportional to the number of chromosomes they contain ( GONZALEZ et al. 1989). In the Harka testes, the diameter of -10% of the onion stage spermatid nuclei was either larger or smaller than those of the regular spermatid nuclei. These spermatids were aneuploid and carried more or fewer chromosomes than the normal ones (Figure 4 ) . It should be mentioned that

TABLE 7 Features of the Hothindwed nondisjunction during spermatogenesis

Males Test period Progeny

Meld of progeny Males tested (days) Total Wild type 6w sp cn bw ed dp cl (progeny/male/d)

mwh e 10 170 1 1 0 0 0 5.9 X 1 0 - ~ Hmka 53 636 1085 35 546 382 122 3.3 x 10-2 Hmka”’ 30 222 3 2 1 0 0 4.5 X

The second chromosome constitution of the males was cn h / e d d p cl. The mwh e, Hmka and Hmka”” males also carried the TM3, S6 Ser chromosome. Single males were mated with C ( 2 ) m , 6w sp females. Hmku”’ stands for the EMS-induced revertant of Hmka. Wild-type flies indicate reductional nondisjunction. The bw sp class results from nullo-2 sperm. The cn bw and ed dp cl classes are indicators of equational nondisjunction. The “test period” is the sum of days throught which the lots were tested.

1594

I

J. Szabad, E. Mithe and J. Pur0

I

"- , I

FIGURE 4.-Cytology of Hmka spermatogenesis. ( A ) Diameter of the "onion" stage spermatid nuclei reflects varying chromosome content (cJ GON7AEL et al. 1989). White and black arrows point to small or large nuclei, with apparently reduced and elevated numbers of chromosomes, respectively. Notice that the two types of nuclei are in close association, indicating that they originate from one spermatocyte through nondisjunction. (B-G) Second meiotic division figures. ( B ) Part of a normal-looking cyst. (C) Part of a cyst with abnormal divisions. (D-G) Enlarged micrographs showing undercondensed chromosomes in meta- or anaphase 11. (The unaffected chromosomes are condensed as shown on B.) ( E ) Trailing and loosely condensed chromosomes are frequently seen between telophase I1 nuclei. (F-G) Early telophase I1 figures with reduced and increased numbers of chromosomes. ( H ) Satellite nuclei, as pointed by the open arrows, were not detected in Hmka cysts, indicating that the chromosomes are not lost in the course of Hmka spermatogenesis. The micrograph on H shows satellite nuclei that derived after loss of a compound autosome during the second division of spermatogenesis (J. PURO, unpublished result). ( I -J) Normal- looking and abnormal sperm bundles from a testis of a Horka male. Notice that the abnormal sperm bundles some sperm heads are not attached to the sperm tails and are abnormally distributed. Scale bars: 10 pm for A and D-H and 20 pm for B, C, I and J.

all the spermatids with reduced diameter were associated Cytological analysis of chromosome content of pri- with a nucleus of increased size, and vice versa, support- mary and secondary spermatocytes confirmed the ge- ing the genetic data: Hmka induces nondisjunction dur- netic data. The first meiotic divisions appeared normal ing spermatogenesis and not loss of chromosomes. in testes of the Harka males. Abnormalities were de-


tected primarily in the second meiotic division. The abnormalities include ( 1 ) abnormal condensation and disarrangement of chromosomes at metaphase, ( 2 ) lagging of chromosome arms that appear to reach the telophase nuclei late and (3) uneven distribution of chromatin at telophase, which is an indication of aneuploid segregation (Figure 4) . It is an important observation that satellite nuclei, which represent lost chromosomes ( PURO 1991 ) , were not detected. Thus, it is very likely that the aneuploid sperm originate through nondisjunction rather than chromosome loss.

The frequency of nullo-2 and diplo-2 sperm was determined as follows. Larvae hatched from 94 (2.2% ) of the 4362 eggs collected from C(2)EN, bw sp females mated with cn bw/ ed dp cl; Hmka/ TM3 (or + ) males. Because nullo-, mono- and trisomy of the second chromosome results in embryonic lethality, the eclosing larvae represent the viable diploid combinations. Viable offspring originate from 50% of the aneuploid sperm, that is, when they randomly combine with C(2)ENde- rived ova. Thus, the frequency of aneuploid sperm is 2 X 2.2% = 4.4%. It should be mentioned that 4.8% of the descending females were gynandromorphs in the C(2)EN experiment.

Aneuploid sperm may, in principle, origmate through nondisjunction during mitotic proliferation of the sper- matogonial cells ( SZABAD 1986). Wild-type and cn bw plus ed d p cl flies should develop with equal frequencies after mitotic nondisjunction. As the results presented in Table 7 contradict the expectations, it is safe to con- clude that Horka does not induce nondisjunction during the mitotic proliferation period of the spermatogo- nial cells.

Many cysts of the secondary spermatocytes appeared to be normal in the Hmka males. However, most if not all the nuclei appeared abnormal in the affected cysts showing an all or none reaction (Figure 4 ) . This observation is related to the fact that spermatids develop side by side in the cytoplasm of a common syncytium ( LINDSLEY and TOKUIASU 1980) . A mutant cytoplasmic Hmka factor thus may exert its effect on every nucleus within the cyst.

Arrangement of the spermatid heads was abnormal in many sperm bundles in the males showing abnormal cysts (Figure 4) . Whether sperm of the abnormal bundles participate in fertilization is not clear. LIFSCHWZ and MEYER ( 1977) and CASTRILLON et al. ( 1993) described recessive male-sterile mutants in which sperm arrangement within the bundles was similar to that seen in abnormal bundles of the Hmka males.

Features of the Horkcbinduced gynandromorphs: To determine the average frequency of maleness (the probability that a given structure is male) , we analyzed 80 Hmkaderived adult gynandromorphs with 42 land- marks each. The average frequency of maleness was 47 ? 27%, and this is in agreement with the hypothesis

that the Hork6derived X chromosome can be lost as early as the gonomeric division. The proportion of the male structures varied between 6% and 94% in the different gynandromorphs, indicating that the Hmka- derived X chromosome can be lost during the gonomeric plus the subsequent cleavage divisions and first during the cleavage divisions. We mention, for comparison, that the average frequency of maleness in gynandromorphs generated by the unstable ring-X chromosome, In(1) wuc, is -46-48% (Table 8) ( HALL et al. 1976; JANNINC 1978; S ~ A D et al. 1979).

The conclusion that the Hmka-derived X chromosomes may indeed be lost during different stages after fertilization is confirmed by analysis of 47 blastoderm stage Horkaderived gynandromorphs. Two- to 4hr-old embryos were collected and fixed from a cross between y v f mal females and Horka/ TM3, Sb Ser males. The female cells were identified with the Sex-lethal antibody using immunohistochemical techniques ( BOPF et al. 1991 ) . Presence of female (staining) cells was apparent in 756 of the 1821 embryos. Seventy-eight (10.3%) of the 756 embryos were gynandromorphs because they were composed of both female and male cells. (Mosaic embryos did not develop among the >2000 embryos that descended from y u f mal females and mwh e/ TM3, Sb Ser males.) The mosaic pattern of 47 gynandromorphs was drawn onto standard grids. The relative size of the female and male patches was measured by planimetry. (The 31 gynandromorphs that developed beyond the blastoderm stage were omitted from the present study.) The average size of the female and male patches was 50 ? 28% and varied between 9% and 90% (Table 8 and Figure 5 ) . Analysis of blastoderm gynandromorphs clearly showed that the Hmka-derived Xchromosome, like the unstable In( 1) wuCring-Xchro- mosome, can be lost ( 1 ) during the gonomeric division (in 13 embryos the ratio of the staining and nonstaining area was -1:l) , ( 2 ) during the gonomeric and the subsequent cleavage divisions (in 18 embryos the nonstaining area included 60-90% of the blastoderm sur- face) and ( 3 ) during the cleavage divisions ( 16 embryos had small nonstaining clones; Figure 5 ) .

For a comparison of gynandromorphs generated by Hmka and the unstable In (1) w"" ring-X chromosome, 80 abdomens of each type of adult gynandromorphs were analyzed under the compound microscope. Distri- bution of the number of male bristles on mosaic hemitergites is shown in Figure 6. The frequency of mosaic hemitergites with more than one yellow bristle /mosaic hemitergite was not significantly different between the two types of gynandromorphs ( P > 0.05, chi-square test; 88 / 800 vs. 11 7 / 800 for Hmka and In (1) w""gynandromorphs, respectively; Figure 6 ) . This result may suggest similar mechanisms for loss of the Hmka exposed and the I n ( l ) wuc X chromosomes. However, in the In( 1) wuc gynandromorphs, 71 of the 800 hemitergites

1596 J. Szabad, E. MgthC and J. Puro

TABLE 8

Proportion of male structures in Horka and Zn(l)wK-generated gynandromorphs

Stage of examination Type of

gynandromorph Blastoderm Larval epidermis Adult cuticle

Horkn 50 ? 28 (47) Not determined 47 ? 27 (80) In(I)rul'(' 34 2 25" (18) 46 2 24h (140) 49 2 29 (80)

Values are means ? SD with number of cases in parentheses. " From Table 2 of ZAIOKAR et a l . 1980. "From SZADAI) el nl. 1979.

carried X 0 clones composed of a single y bristle each (see &POLL 1972). Only four such clones developed on the 800 hemitergites of the Horka-generated gynandromorphs (Figure 6 ) . The majority of the small X 0 clones originated most likely from the loss of the In ( I ) wrIf,' chromosome during proliferation of the his- toblasts, precursors of the tergites ( GARCIA-BELLIDO 1973; GUERRA et al. 1973; SANCHEZ and NOTHIGER 1983). It is also possible that some of the small X 0 y clones originate due to spontaneous mitotic recombination between the In(1 ) w"" and the y v f mal chromosome ( MERIUAM et al. 1972). The absence of small male clones in the Horka-generated gynandromorphs may indicate decay, dilution and/or replacement of the

Horkderived mutant gene product during the course of the cleavage divisions.

DISCUSSION

Horka, a new chromosome destabilizing mutation: Understanding the possible function of the normal H d a + gene product may be facilitated by a comparison of the H d a mutant phenotypes with those characteristic of the other chromosomedestabilizing mutants. There have been three such mutations charac- terized in detail so far: nonclaret disjunctional ( ncd) , which was originally identified by the claret nondisjunctional ( can") alleles; mitotic loss inducer ( mit) and paternal

C .. D

FIGURE 5.-Micrographs of blastoderm-stage XX// X 0 mosaics (gynandromorphs, A-C) and a diplo-2// haplo-2 mosaic ( D ) . The XXcells were made visible by the Sex-lethal antibody ( BOPP et al. 1991 ) and stained dark in a standard histochemical procedure. The male cells do not stain. [Pole cells, founders of the germ line, do not stain even when they are next to female blastoderm cells with whom they share common lineage; cJ: BOPP et al. 19911. Female and male cells cover areas of about the same size when the Horkderived X chromosome is lost during the gonomeric division (A). Male areas can be smaller (B) or larger ( C ) than the female areas and indicate late and successive lost of the X chromosomes, respectively. The diplo-2// haplo- 2 mosaic originated following elimination of a chromosome labeled with the &galactosidase transgene. The diplo-2 cells stained for &galactosidase, whereas the haplo-2 ones remained colorless.


NUMBER OF MALE BRISTLES PER HEMITERGITE

FIGURE 6.-The proportion of male ( y ) bristles on mosaic hemitergites of gynandromorphs generated by Zn(1) w'", the unstable ring-X chromosome, and the dominant paternal- effect mutation Horka. Note the excess of small y clones in the In ( I ) w""-derived gynandromorphs.

loss ( p a l ) (BAKER and HALL 1976; ASHBURNER 1989; LINDSLEY and ZIMM 1992). When exposed to mutant ncd , pal or rnit gametogenesis, the chromosomes are rendered unstable and can be lost in the descending zygotes. The mutations ncd and mit act maternally, and pal acts paternally. In ncd mostly the maternally, in mit both the maternally and the paternally, and in pal only the paternally derived homologs of the zygote chromosomes are affected. Horka is a new chromosome destabilizing mutation. Although ncd , mit and pal are recessive, Hwka is dominant. Chromosome losses take place in the zygotes descending from Horka males; thus Horka possesses a dominant paternal effect. Like in the case of pal, only the paternally derived chromosomes are lost in the Hmk&derived zygotes. The chromosome destabilizing effect of Hmka is reminiscent of chromosome imprinting inasmuch as the paternally derived chromosomes are eliminated ( MOORE and HAIC 1991 ) . (The dominant maternal effect of Horka and the loss-of-function phenotypes will be described elsewhere; J. S Z A B ~ et al., in preparation.)

Like in the case of n c d , mit and pal, the chromosome destabilizing effect of Horka depends on the genetic background. As in ncd, mit and pal, the frequency of X and fourth chromosome loss is usually in the range of 2-4%. However, factors, not readily identifiable, can increase the frequency of X chromosome loss to over 20%. Furthermore, the different chromosomes possess different Horka sensitivity. Although the Horka-exposed X, second, third, fourth and the XY chromosomes can be lost, y+Yand w+Ychromosomes do not respond to Horka action, confirming that the Ychromosome is more stable than either the X or the fourth chromosomes (ASHBURNER 1991). GELB~RT (1974) reported that neither a marked Y nor the XY chromosomes are frequently eliminated by mit.

In contrast to ncd and pal, where loss of the X and

the fourth chromosomes is positively correlated, im- plying that the X 0 and the haplo-4 clones are almost always coincident, loss of the Horku-derived X and fourth chromosomes is not correlated; X 0 cells are not haplo-4 and vice versa.

As in the case of ncd and pal, loss of the Horka- derived Xchromosome can take place during the gonomeric division, as indicated by the production of gynandromorphs with 50% female and 50% male tissues (Figure 5 ) ( BAKER 1974; PORTIN 1978). However, multiple and late Xlosses are also apparent as revealed by gynandromorphs with male clones comprising larger or smaller than 50% of the zygote's tissues, respectively. Horka, like the other chromosome destabilizing mutations and the unstable ring-X chromosome In (1 ) wDC, is a useful tool for the generation of gynandromorphs ( SZABAD and NOTHIGER 1992) as well as diplo // haplo mosaics.

Beside its chromosome destabilizing action, Horka induces nondisjunction, primarily equational, but not chromosome loss during spermatogenesis. (Note that the Horka-related mutant phenotypes stem from the same mutation as shown by the linear relationship between the frequencies of Xchromosome loss and the aneuploid sperm and also the concomitant reversion of these effects.) Of the other chromosome destabilizing mutants, ncd induces nondisjunction and pal induces chromosome loss. ( cJ LINDSLEY and ZIMM 1992 ) .

The possible function of the Hwka protein: It is rather unlikely that the normal Horka+ gene product specifies a component of the spindle apparatus, because only the paternally derived chromosomes are lost in the offspring. In the case of cytoplasmic factors, elimination of both maternally and paternally derived chromosomes would be expected as was described for mit (GELBART 1974). We propose, based on the chromosome destabilizing mutant phenotypes, that the normal Horka+ gene product is required-directly or indirectly-for appropriate chromosome segregation. Because the segregation behavior of the chromosomes resides in the centromere and/or in the neighboring region [as has been elegantly shown based on the segregation patterns of the yeast chromosomes with replaced centromeres ( SIMCHEN and HUCERAT 1993) ], we also propose that the Horka+ gene product is required for normal function of the centromere and/ or nearby chromosome regions.

The Horka mutant phenotype is analogous to those described for ncd: (1 ) an elevated frequency of nondisjunction in the ncd homozygous females, ( 2 ) ncd is maternal in action and ( 3 ) mostly the maternally derived chromosomes are lost in the descending zygotes (DAVIS 1969; SEQUIERA et al. 1989). Horka induces nondisjunction chromosome loss in the zygotes and only chromosomes originating from the mutant

1598 J. Szabad, E. Math6 and J. Puro

parent are lost. Molecular analysis of the ncd gene showed that the ncd gene encodes a kinesin-like minus end-directed microtubule motor protein required for chromosome segregation through association with the kinetochore (YAMAMOTO et al. 1989; ENDOW et al. 1990; MCDONALD and GOLDSTEIN 1990; MCDONALD et al. 1990; ENDOW 1992, 1993; THEURKAUF and HAWLEY 1992). Thus, it appears likely that the normal Horka+ product is required for centromere-associated function.

BAKER ( 1975) proposed, based on the pal-related mutant phenotypes, that the normal pal gene “specifies a product that is component of, or interacts with, the centromeric region of chromosomes.” The chromosomes may have defective centromeric regions in the pal mutant males and can be lost. The molecular nature of the pal product is not know. There are apparent similarities between pal and Horka: both act paternally and only the paternal chromosomes are lost in the zygotes. The similar mutant phenotypes further support the hypothesis that the normal Horka+ product is likely to be required for a centromere-related function.

The gain-of-function nature of Hwku and its rele- vance to the H w h + function: Whether the normal Horka+ product does indeed function in the centromere (or in a nearby chromosome region) cannot be taken for granted. The “traditional” genetic dissection makes use of loss-of-function recessive mutations and deduces the normal gene function from the mutant phenotype, as in the case of, for example, ncd and pal. Horka is a dominant gain-of-function mutation. [It can be reverted and its encoded mutant gene product perdures in the nonmutant germ line cells ( ERDELYI and SZABAD 1989) .] Horka is neomorphic: its dominant paternal effect does not depend on the number of wild-type copies present in the Horka males and the Horka/deficiency combination is lethal (J. SZABAD, unpublished data). Thus, in principle, the mutant Horka product may interfere with another process from that in which the normal Horka+ product is in- volved.

However, results to be described in the forthcoming paragraph support our hypothesis concerning the possible function of the normal Horka+ product. It is apparent that mutant products encoded by all the dominant female-sterile ( F s ) mutations studied thus far disturb the same process in which the normal protein functions. [Recall that Horka is one of the Fs mutations; ERDELYI and SZABAD 1989) .] The dominant as well as the recessive Toll, easter and dorsal Fs alleles interfere with establishment of the dorsoventral embryonic polarity. Although the dominant alleles induce the formation of ventralized and/ or lateralized embryos, dorsalized embryos develop when function of the above genes is eliminated (GOVIND and STEW-

ARD 1991; ST. JOHNSTON and NUSSLEIN-VOLHARD 1992) . The Fs (2 ) torso”4021 allele leads to the formation of “all terminal” embryos ( KLINGLER et al. 1988; SZA- BAD et al. 1989). In contrary, terminal structures do not develop when function of the torso gene product is absent (SCHUPBACH and WIESCHAUS 1986). The rolledsm Fs allele disturbs function of the same signal transducing pathway in which the rolled+ product functions ( BRUNNER et al. 1994). The Tomaj Fs alleles, which identify the so-called maternal alpha-tubulin gene, interfere with microtubule formation (K. JOS VAY, E. MATHE and J. SZABAD, unpublished data). Mi- crotubules do not form in the absence of the maternal alpha-tubulin molecules ( MATTHEWS et al. 1993) . Al- though the data listed above support our hypothesis concerning the function of Horka+ in the centromere and/ or nearby regions, direct evidence will emerge after only molecular cloning of the gene.

We are grateful to Drs. D. BOPP for providing the Sex-lethal antibody; H. BELLEN for the lines with the “all over” expression of the p- galactosidase transgene and MIKLOS ERDELYI, ROLF NOTHIGER, TRUDI SCHOPBACH and an anonymous reviewer for discussions and com- rnentS on the manuscript. The “ Ho-rlza project” was supported by grant no. 922 from the Hungarian National Science Foundation to J. SZABAD and also by the BBstyai-Holczer Memorial Foundation.

Note added in PooJ Complementation analyses with alleles of genes situated around the Hmka locus showed that Horka is allelic to the mutagen-sensitive mutant mus309 (BOYD et al. 1981) and to ck12 (GAUSZ et al. 1981).

LITERATURE CITED

ASHBURNER, M., 1989 Drosophila. A Laboratoly Handbook. Cold Spring Harbour Laboratory Press, Cold Spring Harbor, NY.

BAKER, B. S., 1975 Paternal loss ( p a l ) : a meiotic mutant in Drosophila

267-296. melanogaster causing loss of paternal chromosomes. Genetics 80:

BAKER, B. S., and J. C. HALL, 1976 Meiotic mutants: genetic control of meiotic recombination and chromosome segregation, pp. 351-434 in The Genetics and Biology ofDrosophila, Vol. la, edited by E. NOVITSKI and M. ASHBURNER. Academic Press, New York.

BEI.I.EN, H. J., C. J. O’KANE, C. WILSON, U. GROSSNIKIAUS, R. K. PWON et al., 1989 P-element-mediated enhancer detection: a versatile method to study development in Drosophila. Genes Dev. 3: 1288-1300.

BOPP, D., L. R. BELL, T. W. CLINE and P. SCHEDL, 1991 Develop mental distribution of female-specific Sex-[ethal proteins in Dru- sophila melanogaster. Genes Dev. 5: 403-415.

BOYD, J. B., M. D. GOLINO, K. E. S. SHAW, C. J. OSGOOD and M. M. GREEN, 1981 Thirdchromosome mutagen-sensitive mutants of Drosophila melanogaster. Genetics 97: 607-623.

BRUNNER, D., N. OELLERS, J. SZABAD, W. H. BIGGS 111, S. L. ZIPURSKY et al., 1994 A gain-of-function mutation in Drosophila MAP kinase activates multiple receptor tyrosine kinase signalling path- ways. Cell 76: 875-888.

CASTRILLON, D. H., P. GONCZY, S. ALEXANDER, R. RAWSON, C. G. EBERHART et al., 1993 Toward a molecular genetic analysis of spermatogenesis in Drosophila melanogaster: characterization of male-sterile mutants generated by single Pelement mutagenesis. Genetics 135: 489-505.

DAVIS, D. G., 1969 Chromosome behavior under the influence of claret-nondisjunctional in Drosophila melanogaster. Genetics 61: 577-594.


ENDOW, S. A., 1992 Meiotic chromosome distribution in Drosophila oocytes: roles of two kinesin-related proteins. Chromosome 102: 1-8.

ENDOW, S. A,, 1993 Chromosome distribution, molecular motors and the claret protein. Trends Genet. 9: 52-55.

ENDOW, S. A., S. HEINKOFF and L. SOLER-NIEDZIELA, 1990 Mediation of meiotic and early mitotic chromosome segregation in Dro- sophila by a protein related to kinesin. Nature 345 81-83.

ERD~LYI, M., and J. SZABAD, 1989 Isolation and characterization of dominant female sterile mutations of Drosophila melanogaster. I. Mutations on the third chromosome. Genetics 122: 111-127.

GARCIA-BELLIDO, A,, and J. R. MEW, 1971 Clonal parameters of tergite development in Drosophila. Dev. Biol. 26: 264-276.

GAUSZ, J., H. GWRKOVICS, G. BENCZE, A. A. M. AWAD, J. J. HOLDEN et al., 1981 Genetic characterization of the region between 86F1,2 and 87B15 on chromosome 3 of Drosophila melanogaster. Genetics 9 8 775-789.

GELBART, W. M., 1974 A new mutant controlling mitotic chromosome disjunction in Drosophila melanogaster. Genetics 76: 51-63.

GLOVER, D. M., 1991 Mitosis in the Drosophila embryo-in and out of control. Trends Genet. 7: 125-132.

GONZALEZ, C., J. CASAL and P. RIPOLL, 1989 Relationship between chromosome content and nuclear diameter in early spermatids of Drosophila melanogaster. Genet. Res. 54: 205-212.

GOVIND, S., and R. STEWARD, 1991 Dorsoventral pattern formation in Drosophila: signal transduction and nuclear targeting. Trends Genet. 7: 119-125.

GUERRA, M., J. H. POSTLETHWAIT and H. A. SCHNEIDERMAN, 1973 Development of the imaginal abdomen of Drosophila mlanogaster. Dev. Biol. 32: 361-372.

HALL, J. C., W. M. GELBART and D. R. KANKEL, 1976 Mosaic systems, pp. 265-314 in The Genetics and Biology of Drosophila, edited by M. ASHBURNER and E. NOVITSKI. Academic Press, New York.

HAWLEY, R. S., and W. E. THEURKUF, 1993 Requiem for distributive segregation: achiasmate segregation in Drosophila females. Trends Genet. 9: 310-317.

JANNING, W., 1978 Gynandromorph fate maps in Drosophila, pp. 1-28 in Genetic Mosaics and Cell Dzfferentiation, edited by W. J. GEHRING. Springer-Verlag, New York.

KARR, T. L., 1991 Intracellular sperm/egg interactions in Drosoph- ila: a three-dimensional structural analysis of a paternal product in the developing egg. Mech. Dev. 34: 101-112.

KASTENBAUM, M. A. and K. 0. BOWMAN, 1970 Tables for determining the statistical significance of mutation frequencies. Mutat. Res. 9: 527-549.

KLINGLER, M., M. ERDELYI, J. SZABAD and C. NOSSLEIN-VOLHARD, 1988 Function of tmso in determining the terminal anlagen of the Drosophila embryo. Nature 335: 275-277.

LIFSCHWZ, E., and G. F. MEYER, 1977 Characterisation of male meiotic-sterile mutations in Drosophila melanogaster. Chromosoma 64:

LIN, H., and M. F. WOLFNER, 1991 The Drosophila maternal-effect gene f s(l) Yu encodes a cell cycledependent nuclear envelope component required for embryonic mitosis. Cell 6 4 49-62.

LINDSLEY, D. L., and K T. TOKWMU, 1980 Spermatogenesis, pp. 226-294 in The Genetics and Biology of Drosophila, Vol. 2d, edited by M. ASHBURNER and T. R. F. WRIGHT. Academic Press, New York.

LINDSLEY, D. L., and G. G. ZIMM, 1992 Thz Genome of Drosophila mlanogaster. Academic Press, San Diego and London.

M A ~ H E W S , K. A,, D. REES and T. C. KAUFMAN, 1993 A functionally specialized a-tubulin is required for oocyte meiosis and cleavage mitoses in Drosophila. Development 117: 977-991.

MCDONALD, H. B., and L. S. B. GOLDSTEIN, 1990 Identification and characterization of a gene encoding a kinesin-like protein in Drosophila. Cell 61: 991-1000.

MCDONALD, H. B., R. J. STEWART and L. S. B. GOLDSTEIN, 1990 The kinesin-like ncd protein of Drosophila is a minus enddirected microtubule motor. Cell 63: 1159- 1165.

371-392.

MEW, J. R., R. NOTHIGER and A. GARCIA-BELLIDO, 1972 Are di- centric anaphase bridges formed by somatic recombination in X chromosome inversion heterozygotes of Drosophila melanogasterr? Mol. Gen. Genet. 115 294-301.

MOORE, T., and D. HAIG, 1991 Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 7: 45-49.

PORTIN, P., 1978 Studies on gynandromorphs induced with the claret-nondisjunctional mutation of Drosophila melanogaster. An approach to the timing of chromosome loss in cleavage mitoses. Heredity 41: 193-203.

PURO, J., 1991 Differential mechanisms governing segregation of a univalent in oocytes and spermatocytes of Drosophila melanogaster. Chromosoma 100: 305-314.

PURO, J., and S. NOKKAIA, 1977 Meiotic segregation of chromosomes in Drosophila mlanogaster oocytes. A cytological approach. Chromosoma 63: 273-286.

RIPOLL, P, 1972 The embryonic organization of the imaginal wing disc of Drosophila melanogaster. Wilhelm Roux' Arch. 169: 200- 215.

SANCHEZ, L., and R. NOTHIGER, 1983 Sex determination and dosage compensation in Drosophila melanogaster: production of male clones in XX females. EMBO J. 2 485-491.

SCH~IPBACH, T., and E. WIESCHAUS, 1986 Maternaleffect mutations altering anterior-posterior pattern of the Drosophila embryo. Roux's Arch. Dev. Biol. 195: 302-317.

SCH~PBACH, T., E. WIESCHAUS and R. NOTHIGER, 1978 A study of the female germ line in mosaics of Drosophila. Roux's Arch. Dev. Biol. 184: 41-56.

SEQUEIRA, W., C. R. NELSON and P. SZAUTER, 1989 Genetic analysis of the claret locus of Drosophila melanogaster. Genetics 123 51 1 - 524.

SIMCHEN, G., and Y. HUGERAT 1993 What determines whether chromosomes segregate reductionally or equationally in meiosis? Bio- Essays 15: 1-8.

ST. JOHNSTON, D., and C. NUSSLEIN-VOLHARD, 1992 The origin of pattern and polarity in the Drosophila embryo. Cell 6 8 201- 219.

SZABAD, J., 1978 Quick preparation of Drosophila for microscopic analysis. Dros. Inf. Serv. 53: 215.

SZABAD, J., 1986 A genetic assay for the detection of aneuploidy in the germ-line cells of Drosophila melanogaster. Mutat. Res. 164:

SZABAD, J., and R. NOTHIGER, 1992 Gynandromorphs of Drosophila suggest one common primordium for the somatic cells of the female and male gonads in the region of abdominal segments 4 and 5. Development 115: 527-533.

SZABAD, J., T. SCHUPBACH and E. WIESCHAUS, 1979 Cell lineage and development in the larval epidermis of Drosophila melanogaster. Dev. Biol. 73: 256-271.

SZABAD, J., M. ERDELYI, G. HOFFMANN, J. SZIDONVA and T. R. F. WRIGHT, 1989 solation and characterization of dominant female sterile mutations of Drosophila mlanogaster. 11. Mutations on the second chromosome. Genetics 122 823-835.

THEURKAUF W. E., and R. S. HAWLEY, 1992 Meiotic spindle assembly in Drosophila females: behaviour of nonexchange chromosomes and the effects of mutations in nod kinesin-like protein. J. Cell Biol. 116: 1167-1180.

WALD H., 1936 Cytologic studies on the abnormal development of the eggs of the claret mutant type of Drosophila simulans. Genetics 21: 264-281.

WRIGHT, T. R. F., 1970 The genetics of embryogenesis in Drosoph- ila. Adv. Genet. 15: 261-395.

YA"OT0, A. H., D. J. KOMMA, C. D. SHAFFER, V. PIRROTTA and S. A. ENDOW, 1989 The claret locus in Drosophila encodes products

305-326.

required for eyecolor and for meiotic chromosome segregation. EMBO 1. 8 3543-3552.

ZALOKAR, M., I. ERK and P. SANTAMARIA, 1980 Distribution of ring- X chromosomes in the blastoderm of gynandromorphic D. melanogaster. Cell 19: 133-141.

Communicating editor: T. SCH~PBACH

hurku, a dominant mutation of drosophila, induces nondisjunction

Documents