The Tomaj mutant alleles of αtubulin67c reveal a requirement for the encoded maternal specific tubulin isoform in the sperm aster, the cleavage spindle apparatus and neurogenesis during embryonic development in Drosophila

Initiation of embryogenesis in Drosophila requires the contribution of both maternal and paternal components, e.g. the female and male pronuclei, the egg cytoplasmic factors produced during oogenesis and the sperm-derived centriole. A set of intricate events, shortly after meiosis I metaphase release and fertilization lead to commencement of embryogenesis (Foe et al., 1993; Callaini and Riparbelli, 1996; Riparbelli and Callaini, 1996; Sagata, 1996). We have previously described a collection of dominant female sterile (Fs) mutations that, based on their phenotypes, appear to identify important and/or specific components required for the initiation of embryogenesis in Drosophila (Erdélyi and Szabad, 1989; Szabad et al., 1989).

This paper presents the developmental genetic and molecular analysis of three dominant female sterile Tomaj (Tomaj^D1-3) alleles and their eleven revertant (Tomaj^R) alleles that identify the αTub67C locus, a member of the Drosophila α-tubulin gene family. This gene family consists of four genes: αTub67C, αTub84B, αTub84E and αTub84D (Kalfayan and Wensink, 1981; Fryberg and Goldstein, 1990). The αTub84B and the αTub84D genes are constitutively expressed during all stages of development and encode isotypes differing at only 2 of 450 amino acids (Kalfayan and Wensink, 1982; Theurkauf et al., 1986; Matthews et al., 1989). The αTub84E encoded protein differs from the constitutively expressed isotypes at 21 of 450 amino acids and is expressed only in the testis and peripheral nervous system (Theurkauf et al., 1986; Bo and Wensink, 1989; Matthews et al., 1990). The αTub67C gene encodes the most divergent α-tubulin sharing only 65% amino acid identity with the constitutively expressed α-tubulins (Theurkauf et al., 1986). The αTub67C gene is transcribed only in the female germ line during oogenesis, and the encoded mRNA is present in the ovaries and in early embryos (Matthews et al., 1989). Translation occurs only in the ovaries, and the αTub67C isotype accounts for 20% of the α-tubulin pool in a mature egg (Kalfayan and Wensink, 1982; Natzle and McCarthy, 1984; Matthews et al., 1989, 1990, 1993).

The α- and β-tubulin heterodimer represents the structural unit of microtubules (MTs). In general, MTs are important cytoskeletal proteins with different functions during both the cell cycle and development. Structural and functional heterogeneity of the MTs derives from the α- and β-tubulin heterodimers as well as the MT associated proteins (MAPs) e.g. motor molecules and stabilizing proteins. Alternatively, the MAPs can act as targets for intracellular regulatory signals (Holzbaur and Valle, 1994; Moore and Endow, 1996). Furthermore, different MT functions may originate from the combination of numerous posttranslational modifications like tyrosination-detyrosination, polyglycylation, polyglutamylation, acetylation and phosphorylation-dephosphorylation (Joshi and Cleveland, 1990; Bulinski and Gundersen, 1991; Theurkauf, 1992; Paturle-Lafanechere et al., 1994; Larcher et al., 1996; Mary et al., 1996; Multigner et al., 1996). Additionally, modulation of MT activity during the cell cycle is governed by centrosomes. Centrosomes play key roles in the nucleation of MT polymerization, MT depolymerization, release and capture of MTs as well as the establishment of MT polarity (Kimble and Kuriyama, 1992; Schatten, 1994; Raff, 1996).

During Drosophila embryogenesis, the αTub67C isotype seems to have a specialized or restricted functional capacity. Matthews et al. (1993) and Komma and Endow (1997) have proposed, based on phenotypes associated with the αTub67C^1-4 mutant alleles, that the αTub67C isotype is required for the proper structure and function of the MTs during female meiosis and cleavage divisions. Further, the possible involvement of the αTub67C isotype in differentiation of the embryonic nervous system was proposed by Theurkauf (1992).

The Tomaj^D and Tomaj^R mutant alleles provide new insights into the function of the αTub67C gene during Drosophila embryonic development. Our developmental genetic analyses have shown that the function of the αTub67C gene is required for: (1) the formation of a prominent sperm aster, which ensures the appropriate positioning of the male and female pronuclei; (2) the normal structure and function of the cleavage spindle apparatus; and (3) the differentiation of the embryonic nervous system, which requires MTs that contain the αTub67C tubulin isotype.

Drosophila cultures were kept at 25°C on standard cornmeal agar medium. Marker mutations, the Dp(3;3)S2a2 duplication and the Df(3L)AC1 deficiency as well as the balancer chromosomes were described by Lindsley and Zimm (1992). The three female sterile Tomaj^D mutations and their EMS-induced revertants were previously described by Erdélyi and Szabad (1989), and are summarized in Table 1. Additional revertants were induced by P-element and X-ray mutagenesis, according to the method of Erdélyi and Szabad (1989). We also made use of the αTub67C^1-4 mutant alleles that were described by Matthews et al. (1993).

For the developmental genetic analysis of the mutant phenotypes, egg chambers and eggs derived from females with the following genotypes were analyzed: wild-type (Canton-S), Tomaj^D/+, Tomaj^D/+/+ (where + stands for the Dp(3;3)S2a2 duplication), TomajD/- (where -stands for the Df(3L)ACl deficiency), Tomaj^D/Tomaj^R, Tomaj^R/Tomaj^R, Tomaj^R/αTub67C^1-4. Females with the above mentioned genotypes were mated with wild-type males and over a hundred of their eggs were collected and analyzed. In some experiments we also analyzed eggs of virgin Tomaj^D/+, Tomaj^R/-and wild-type females.

To analyze meiosis I, mature oocytes were isolated and fixed according to the method of Theurkauf and Hawley (1992) and Puro (1991). For analysis of meiosis II and the pronuclear stage of embryogenesis, eggs were squeezed out from the uterus. Whole-mount eggs and embryos were processed and fixed according to Whitfield et al. (1988), Matthews et al. (1990), Karr (1991), Puro (1991) and Maldonado-Codina and Glover (1992). The Feulgen-Giemsa double stainings were carried out according to the method of Puro (1991).

The MTs were detected with YL1/2 rat monoclonal anti-tubulin (Sera Lab) or mouse monoclonal anti-αtubulin (Amersham) antibodies. Centrosomes were detected by either the Rb188 rabbit anti-centrosome (specific for CP190 antigen) or CNN rabbit anti-centrosomin antibodies (Whitfield et al., 1988; Li and Kaufman, 1996). The mouse T47 anti-lamin monoclonal antibody was used to stain the nuclear lamina (Paddy et al., 1990). The sperm tail was detected with the Ax-D5 mouse monoclonal antibody (Karr 1991). The embryonic central and peripheral nervous systems (CNS and PNS) were stained with the 22C10 mouse monoclonal antibody (Fujita et al., 1982; Goodman and Doe, 1993; Jan and Jan, 1993).

Samples were incubated in the primary antibodies overnight at 4°C. The secondary antibodies were obtained from Jackson Immunoresearch Laboratories Inc. (Avondale, PA). Incubations with the secondary antibodies were performed either at room temperature for 2 hours or overnight at 4°C. When appropriate, DNA was stained with propidium iodide (PI). Specimens were mounted in 85% glycerol, containing 2.5% n-propyl gallate.

A Zeiss Axiolab light microscope was used to analyze the Feulgen-Giemsa stained specimens. Digital images of optical sections through fixed and immunostained specimen were collected with a Zeiss 410 confocal laser scanning microscope. Confocal images were analyzed by the Carl Zeiss Microscope System LSM version 3.59 software.

The Tomaj^D2 allele is temperature sensitive. At 18°C, Tomaj^D2 has a fully penetrant dominant female sterile phenotype. However, at 29°C larvae hatch from about 20% of the eggs, most of which develop into adults. The remaining 80% of the eggs show polyphasic lethality. To determine the temperature-sensitive period, Tomaj^D2/+ females were mated to Canton-S males and temperature shift experiments performed on the crosses. In the shift-down experiment, flies were shifted down from 29°C to 18°C, eggs collected at two hour intervals and the hatch rate monitored. In the shift-up experiment, the parent flies were transferred from 18°C to 29°C, eggs collected at one hour intervals and the hatch rate determined. Preparations of the cuticle-producing dead embryos were analyzed to identify their mutant phenotype (Wieschaus and Nüsslein-Volhard, 1986).

To decide whether the Tomaj^R mutations affect the female germ line or soma, two types of germ line chimeras were constructed. In the first set of experiments Tomaj^RE1.1/Tomaj^RE3.1, Tomaj^RE2.1/Tomaj^RE3.1 pole cells (progenitors of germ line cells) were transplanted into Fs(1)K1237/+ host females (Table 2). The Fs(1)K1237 (= ovo^D1) mutation blocks germ line function without affecting the somatic cells (Komitopoulou et al., 1983). The eclosing Fs(1)K1237/+ host females were mated with mwh e males and the progeny scored for these markers. In the reciprocal experiment wild-type pole cells marked with y v f mal marker mutations were transplanted into Tomaj^RE1.1/Tomaj^RE3.1, Tomaj^RE2.1/Tomaj^RE3.1 host embryos (Table 3). The eclosing Tomaj^R/Tomaj^R host females were mated with yvfmal males and the progeny scored for these markers. The eclosing Tomaj^R/TM3, Tomaj^R/TM6 and TM3/TM6 sibling chimeras served as internal controls.

The 7 kb BamHI clone containing the wild-type αTub67C gene was kindly provided by K. A. Matthews (Kalfayan and Wensink, 1981; Matthews et al., 1989). Since the αTub67C gene is included in a BamHI fragment, genomic libraries were generated after partial BamHI digestion from Canton-S, mwh e as well as each of the Tomaj^D/Df(3L)AC1 flies. The αTub67C clones were identified in two steps: first by an α-tubulin specific 1.7 kb XhoI-EcoRI probe and by a second 96 bp BstNI-AvaII probe specific only for the αTub67C gene. The five clones containing the two wild-type and the three Tomaj^D alleles were subcloned into M13 and then were sequenced on both strands by the dideoxy method using the USB DNA Sequencing Kit (Sequenase Version 2.0).

For the bacterial expression of wild-type αTub67C and the three Tomaj^D encoded αTub67C isotypes, DNA fragments corresponding to the cDNA region were generated by PCR both from the cloned and the genomic DNAs. Two primers were used: 5’CGTA CAT ATG CGA GAA GTA GTG TCC, and 5’GCTA GGA TCC TTA GAA CTC ATC GAA GTC C. The PCR fragments were cleaved with NdeI and BamHI at the primer region and inserted into pET3d expression vectors (Promega), resulting in pETαTub67C⁺, pETTomaj^D1, pETTomaj^D2 and pETTomaj^D3 plasmids. The Escherichia coli BL21 cells transformed with the expression constructs produced the expected proteins following IPTG induction. The bacterially expressed proteins were identified on western blots probed with a polyclonal antiserum against the αTub67C isotype.

To generate polyclonal antiserum against αTub67C, an oligopeptide was synthesized corresponding to amino acid residues 32-61. This sequence is specific to the αTub67C isotype among the Drosophila α-tubulins. The serum was affinity purified and used in western blot analyses. On western blots of bacterially expressed proteins, the antiserum solely identified the wild-type and Tomaj^D1-3 encoded Tub67C isotypes in the same molecular mass range (48 kDa). Unfortunately, the antiserum does not recognize αTub67C in situ (following different fixation procedures), suggesting that the antigenic determinants within the N-terminal domain of αTub67C tubulin are unexposed on the MT surface or were masked by other proteins.

Heterozygous or hemizygous Tomaj^D1 and Tomaj^D3 females are sterile: they lay normal numbers of phenotypically normal eggs that do not hatch. Sperm tail and centrosome antibody stainings clearly show that the majority of the eggs are fertilized. Typically, the first cleavage (gonomeric) division is never established in these eggs (Fig. 1).

To elucidate whether abnormal female meiosis is the reason for the absence of the gonomeric division, we analyzed stage 13 and 14 oocytes and eggs from the uteri of wild-type and both Tomaj^D1 and Tomaj^D3 heterozygous and hemizygous females. The oogenetic stages 13 and 14 correspond to meiosis I prophase and metaphase, while a high proportion of the eggs squeezed out of the uterus show meiosis II. It was apparent, based on both Feulgen-Giemsa and combined anti-tubulin and DNA stainings, that both meiosis I and II proceeded normally (data not shown), and thus defective female meiosis is unlikely to be the source of the lack of commencement of embryogenesis in eggs of Tomaj^D1 or Tomaj^D3 heterozygous or hemizygous females.

The anti-tubulin, anti-centrosome (Rb-188, CNN), anti-lamin and DNA stainings on the newly deposited eggs by heterozygous and hemizygous Tomaj^D1 and Tomaj^D3 females revealed that the primary developmental defect seen was associated with sperm aster structure and function. In wild-type eggs, a small sperm aster emanates from the reorganized and replicated-separated centrosome after fertilization (Fig. 1A). The small sperm aster is composed of MT arrays and remains small until meiosis II anaphase. It then enlarges and becomes prominent by the end of meiotic II telophase (Callaini and Riparbelli, 1996) (Fig. 1B). In fertilized eggs of heterozygous and hemizygous Tomaj^D1 and Tomaj^D3 females a small sperm aster forms similar to that seen in wild-type fertilized eggs (Fig. 1C). However, the small sperm aster never grew to a prominent sperm aster in over a thousand Tomaj^D1 or Tomaj^D3 derived eggs. Failure of prominent sperm aster formation appears thus to be the primary Tomaj^D1 and Tomaj^D3 associated defect. Additionally, it appears that in the absence of a prominent sperm aster the subsequent migration of the female pronucleus towards the male pronucleus does not take place, and thus embryogenesis is blocked prior to the gonomeric division.

It is also interesting that in wild-type eggs, prominent sperm aster formation occurs after the first embryonic centrosome reorganization, replication and separation. The accumulation of CP190 and centrosomin in the first embryonic centrosome as a consequence of the reorganization of the paternally inherited basal body was evident in eggs of wild-type and both Tomaj^D1 and Tomaj^D3 females (Fig. 1A,C). As a rule, the sperm asters remained small, and the first embryonic centrosome clearly replicated once or twice, but the daughter centrosomes failed to separate in the Tomaj^D1 or Tomaj^D3 derived eggs (Fig. 1C).

Concomitant with the failure of prominent sperm aster formation, the polar bodies migrate toward each other by means of MTs and fuse: first three and subsequently all four of the polar bodies combine (triplo- and tetra-fusion) (Fig. 2). The triplo-or tetra-fusion suggests that in the absence of female pronucleus specification and migration all four female meiotic products will differentiate into polar bodies. The chromatin decondensation inside the four polar body nuclei and the sperm head/male pronucleus takes place synchronously, indicating the normal functioning of regulatory mechanisms that synchronize pronuclear behaviour. The observation that triplo- and tetra-fusions are the common rule in unfertilized eggs of both wild-type and heterozygous and hemizygous TomajD¹ and TomajD³ females (approx. 500 and over a thousand analyzed eggs, respectively), implicates the involvement of the prominent sperm aster in the specification of female pronucleus and its migration towards male pronucleus.

TomajD²/+ and TomajD²/-females are sterile at 18°C and the TomajD¹ or TomajD³ phenotype noted above is seen in about 10% of their eggs. In the remaining 90%, embryogenesis proceeded to stages 13 to 17, but was not completed. At 29°C, however, about 20% of the eggs hatch and a significant fraction of the larvae develop to adulthood. Thus, the TomajD² alleles possesses dominant cold-sensitive female sterility, and temperature shift experiments were used to determine the time(s) that the aTub67C gene product is required. In the shift-up experiment the TomajD2/+ females were raised to 18°C and then shifted to 29°C. One hour after the shift up, the 0% hatch rate elevated to 20% and subsequently remained at 20%. In the shift-down experiment the TomajD²/+ females were raised to 29°C and then shifted to 18°C. The hatch rate dropped from 20% to 1% two hours after the shift down. These results show that the TomajD² associated temperature-sensitive period covers stage 14 of oogenesis and the pronuclear stage of embryonic development.

At both 18°C and 29°C, 90% and 70%, respectively, of the TomajD²/+ derived eggs escaped from pronuclear arrest and embryogenesis proceeded to stages 13 to 17, but was not completed. In such embryos the cleavage divisions, blastoderm formation and gastrulation proceeded normally as was inferred by immunostainings with anti-tubulin, anti-lamin antibodies and DNA staining (data not shown). Although these embryos do not complete embryogenesis, they produce some cuticular structures. Since the cuticle of the dead embryos was reminiscent of the so-called neurogenic phenotype (Campos-Ortega, 1993), we analyzed both CNS and PNS of the TomajD2 derived embryos by immunostaining (Fig. 3). Observation of the CNS and PNS showed highly disorganized neuronal structures. The brains remained rudimentary, while the commissural and longitudinal processes of the ventral nerve cord were disrupted. Interruptions in the PNS were also apparent, e.g. the sensory neurons and their axons were abnormally oriented. Theurkauf (1992) has shown that the maternally supplied αTub67C isotype perdures into the embryonic nervous system. The phenotype of the escaper animals produced by the TomajD2 allele now demonstrates a requirement for this isotype in formation of the embryonic CNS and PNS.

To determine whether the TomajD^1-3 encoded mutant αTub67C isotypes participate in the same events as the wild-type isotype or interferes with normal processes, we constructed TomajD/+/+ females, where + stands for the Dp(3;3)S2a2 duplication which covers the αTub67C locus. The TomajD/+/+ females were fertile: 26%, 5% and 8% of the eggs developed to adults in the case of TomajD¹, TomajD² and TomajD³, respectively. Since an extra copy of the wild-type locus is capable at least partially of ameliorating the defects associated with the dominant alleles, it would appear that the TomajD^1- alleles are antimorphic and their products disrupt or poison the same phenomena in which the normal αTub67C isotype functions.

To determine the molecular nature of the gain-of-function TomajD mutations, we cloned and sequenced the TomajD^1-3 alleles. Single base pair substitutions in two different positions resulted in replacement of Glu^-Lys⁺ in both the TomajD¹ and TomajD³ encoded αTub67C isotypes, while in the TomajD² encoded product a Gln/Arg⁺ change has occurred (Table 4). We also expressed the wild-type as well as the TomajD^1-3 alleles in bacteria and established using western blots probed with a polyclonal serum against αTub67C, that the TomajD^1-3 alleles encode a normal-sized peptide (data not shown). On the basis of these data we can conclude that the gain-of-function TomajD^1-3 mutations lead to single amino acid changes that may identify important functional sites in αTub67C.

To determine the loss-of-function Tomaj phenotype, we reverted the TomajD^1-3 alleles by EMS, X-ray and P-element mutagenesis (Table 1). In situ hybridizations to the polytene chromosomes of the P-induced revertants identify a common insertion point in 67C4-5, where the αTub67C^1-4 mutant alleles were previously mapped (Matthews et al., 1993). The αTub67C^1-4 mutant alleles fail to complement the Tomaj^R alleles, and therefore the two groups of mutations identify the same locus.

The Tomaj^R/-, heterozygous Tomaj^R/Tomaj^R, Tomaj^D/Tomaj^R and Tomaj^R/αTub67C^1-4 combinations are all female sterile. The female sterility of αTub67C^1-4/- hemizygous combinations has already been described by Matthews et al. (1993). Based on the extent of embryogenesis in eggs of the Tomaj^R/- females, the Tomaj^R alleles can be ranked from the strongest to the weakest alleles as follows: RE3.1=RE3.2=RE3.3>RE2.1=RX2.2=RX2.1=RE1.1=RX3.1>R X1.1=RP3.1=RP1.1.

Meiosis I and II proceeded normally in oocytes and eggs of Tomaj^R/-, Tomaj^R/Tomaj^R, Tomaj^D/Tomaj^R and Tomaj^R/αTub67C^1-4 combinations (where R stands for RE3.1, RE3.2, RE2.1, RE1.1, RP1.1) as was shown by anti-tubulin and DNA stainings and Feulgen-Giemsa staining (approx. 500 eggs/combination were analyzed).

Only a small sperm aster formed in variable frequencies (20-70%), in the fertilized eggs of Tomaj^R/-females (where R stands for RE3.1, RE3.2, RE3.3, RE2.1, RE1.1, RX2.1, RX2.2). There is no apparent female pronuclear specification nor is there evidence of migration toward the male pronucleus. Additionally, triplo-or tetra-fusion of polar bodies is common. Thus, the gain-of-function and loss-of-function Tomaj phenotypes are almost identical, implying that in eggs of Tomaj^D1-3 females the encoded mutant αTub67C molecules block the function(s) of the normal αTub67C ones.

The cleavage divisions commenced and proceeded through a variable number of cycles in 20-60% of the eggs laid by females hemizygous and heterozygous for these combinations of the Tomaj^R alleles. The gonomeric division is not completed in eggs of hemizygous and heterozygous females for stronger Tomaj^R alleles e.g. RE3.1, RE3.2, RE3.3, RE2.1, RE1.1. Embryogenesis comes to a cessation before or shortly after cortical migration of cleavage nuclei in eggs of females heterozygous for weaker (leaky) Tomaj^R alleles e.g. RE1.1, RX2.1, RX2.2, RP1.1, RP3.1. The αTub67C^1-4alleles are weak hypomorphs (Matthews et al., 1993). Consistent with this fact the cleavage divisions progressed up to blastoderm formation in eggs of the Tomaj^RE1.1/αTub67C^1-4 females.

The most commonly observed cleavage spindle apparatus defect is related to the gonomeric spindle (Fig. 4A,B). In eggs showing an abnormal gonomeric spindle, the female and male pronuclei approach each other presumably using a functional sperm aster, but they never fuse. In wild-type eggs the gonomeric spindle has well defined poles marked by centrosomes and the convergence of aster MTs. In eggs of Tomaj^R/- and heterozygous Tomaj^R/Tomaj^R females, a barrel-shaped nonfunctional gonomeric spindle appears with apparently greater than normal accumulation of centrosomin at the presumptive poles. Another characteristic feature of the defective cleavage spindle apparatus was the gradual appearance of abnormal ectopic cleavage spindles. This abnormality becomes more severe with time (Fig. 4D-H). Barrel-shaped, undersized, mono- and multipolar spindles as well as fused spindles and patches of MTs developed among the wild-type looking cleavage spindles, often in ‘clonal’ fashion. Replicated, but poorly separated and free centrosomes were seen in close vicinity of MT patches. In degenerating embryos MT patches without centrosomes and chromosomes are rather common.

To determine whether the αTub67C gene is required in the female germ line and/or in soma, we constructed two types of germ line chimeras. First, Tomaj^RE1.1/Tomaj^RE3.1 and Tomaj^RE2.1/Tomaj^RE3.1 germ line cells were transplanted into Fs(1)K1237/+ hosts that have a normal soma and nonfunctional germ line (Table 2). All 15 female chimeras were sterile and possessed the same mutant phenotype as the donor females. In the second experiment normal (y v f mal) pole cells were transplanted into Tomaj^RE1.1/Tomaj^RE3.1 and Tomaj^RE2.1/Tomaj^RE3.1 hosts (Table 3). The 14 chimeric Tomaj^R/Tomaj^R host females in addition to their own non-hatching eggs, produced y v f mal offspring. Thus, it would appear that the αTub67C gene functions only in the female germ line.

The Tomaj mutants map to the αTub67C gene. In Drosophila, early embryogenesis is governed by the maternal dowry, that is, females provide a supply of factors in the egg cytoplasm, including the α-tubulins. The function(s) of the αTub67C gene appear(s) to be restricted to late oogenesis, early embryogenesis and the development of the embryonic nervous system (Komma and Endow, 1995, 1997; Matthews et al., 1990, 1993; Theurkauf et al., 1992; Theurkauf, 1994). As we report in the present paper, the gain-of-function Tomaj^D1-3 and the loss-of-function Tomaj^R revertant alleles provide additional information about the functions of the αTub67C gene.

Genetic dissection of normal gene function is usually based on the phenotype caused by null and/or hypomorphic (leaky) alleles. As we have shown, the Tomaj^D alleles are antimorphic and this fact could obscure any interpretation of αTub67C function based on the phenotype associated with these dominant alleles. However, the phenotype of the strong Tomaj^R/- combinations is almost identical with the one seen in Tomaj^D1 and Tomaj^D3. Thus, it would appear that the Tomaj^D encoded mutant peptides participate in the same processes as the wild type, and consequently the corresponding mutant phenotypes reveal the normal αTub67C function(s). Moreover, since the Tomaj^D/+ and Tomaj^D/- phenotypes are essentially identical, it is clear that the Tomaj^D encoded protein nullifies the wild-type function of the αTub67C isotype and in this respect the dominant alleles act like nulls. This conclusion is supported by the fact that the ‘poison’ effect can be partially titrated out by additional copies of the wild-type αTub67C locus. The apparent nullifying activity of the Tomaj^D1-3 alleles may be explained by their associated molecular lesions. In both Tomaj^D1 and Tomaj^D3 (in different positions) there is a Glu^-/Lys’ missense change, while in the TomajD² encoded tubulin a Gln/Arg+ alteration is found (Table 4). The α-tubulin molecules form heterodimers with β-tubulins and comprise the major constituents of MTs. It would appear therefore that the Tomaj^D1-3 encoded mutant αTub67C molecules form heterodimers with the available β-tubulins and these are incorporated into MTs. Once incorporated into MTs, elongation is apparently blocked, e.g. there is no transition from the small to the prominent sperm aster (see below). In this model the failure of MT elongation is the consequence of tubulin heterodimers that are defective due to the Glu^-/Lys+ and Gln/Arg+ changes in the mutant αTub67C isotypes. Several possibilities could account for the observed defects e.g. an alteration in charge affecting the conformation of peptide, a defective MAP association site or aberrant posttranslational modification sites. Determining which if any of these may be correct will require further analysis.

αTub67C is required for sperm aster formation and function. Since early embryogenesis does not commence in eggs of the Tomaj^D1, Tomaj^D3 and in females hemizygous for stronger Tomaj^R alleles (RE3.1, RE3.2, RE3.3, RE2.1, RE1.1), it appears that the Tomaj^D1, Tomaj^D3 and the stronger Tomaj^R alleles identify an αTub67C function, which is required at or just subsequent to fertilization. However, the αTub67C function is required during female meiosis as was proposed by Matthews et al. (1993), Matthies et al. (1996) and Komma and Endow (1997), based on the αTub67C^1-4 alleles and the use of genetic assays for the detection of nondisjunction and/or chromosome loss. We did not detect abnormal meiotic figures using immunocytological or Feulgen-Giemsa staining procedures in oocytes of Tomaj^D1-3 and different combinations of Tomaj^R females. The difference between the former and present results can be rationalized by assuming that: (1) αTub67C^1-4 and Tomaj^D (and their revertants) identify different αTub67C functions. (2) The cytological methods are much less sensitive to detect abnormal chromosome segregation than the genetic approaches (Szabad, 1986). The Tomaj^D alleles being sterile, cannot be used in genetic tests for analysis of chromosome segregation.

The first observable Tomaj^D1, Tomaj^D3 and Tomaj^R associated defects appear after fertilization when only a small sperm aster is developed (Fig. 1). This compares to wild-type eggs in which the small sperm aster grows rapidly into the prominent sperm aster. We propose, based on the Tomaj mutant phenotype, that the αTub67C isotype is required for formation and function of the prominent sperm aster. Furthermore because the size of the small sperm aster is basically identical in all the mutant genotypes studied and persists for about two hours, we conclude that the αTub67C isotype is not required for small sperm aster formation.

Failure of prominent sperm aster formation leads to the absence of female pronucleus specification and migration toward the male pronucleus. The four meiotic products of female meiosis appear to be equivalent and frequently all four fuse (tetra-fusion). Similarly, Komma and Endow (1997) and Williams et al. (1997) also reported defective female pronuclear formation and fusion of the four polar body nuclei in eggs of the ncd and klp3a mutant females, suggesting that the NCD and KLP3A kinesin-like motor molecules are also required for female pronuclear migration. We propose that the NCD and KLP3A ensure female pronuclear migration along the prominent sperm aster MTs.

αTub67C is required for cleavage spindle apparatus. Cleavage spindle apparatus defects appeared in 20-60% of the eggs derived from different combinations of weaker Tomaj^R alleles (RE1.1, RX2.1, RX2.2, RP1.1, RP3.1), (Fig. 4). The most obvious cleavage spindle defect is the nonfunctional, barrel-shaped gonomeric spindle, which presents excessive centrosomin protein accumulation at the presumptive poles. In embryos that pass the gonomeric division, the embryogenesis ceases at the time of cortical migration, showing highly aberrant cleavage spindles with abnormal centrosome and chromosome displacement. Based on the Tomaj^R associated phenotypes we conclude that the normal αTub67C is specifically needed for the proper structure and function of cleavage spindle apparatus. Whether the αTub67C allows the unusually rapid assembly and disassembly of Drosophila cleavage spindles as was proposed by Matthews et al. (1993) or influences the nucleation of MTs by centrosomes remains to be answered.

αTub67C is required in the embryonic nervous system. At permissive temperature embryogenesis proceeds up to gastrulation in eggs of the Tomaj^D2/+ females. However, at this point most of the embryos die presumably due to the ‘poisonous’ maternal dowry. The dead embryos show abnormal development of both CNS and PNS (Fig. 3). Based on perdurance of the maternally loaded αTub67C isotype in the CNS, Theurkauf (1992) proposed the protein showed CNS-specific developmental stability. Although the nature of developmental stability of αTub67C is not known, the neural defects observed in Tomaj^D2/+ derived embryos indicate that its presence must have functional consequences in the developing embryonic nervous system that may lead to formation of MT bundles in cell division or other neuronal processes.

We thank Miklós Erdélyi for comments on the manuscript and Kathleen Matthews for providing the αTub67C specific clones and mutants. We are grateful to Tim Karr for the Ax-D5 sperm tail-specific antibody, William Whitfield for Rb188 centrosome-specific antibody and Harold Saumweber for the T47 lamin-specific antibody. We also acknowledge the excellent technical assistance of Csilla Papdi. Support for this work came from several sources: Hungarian National Science Foundation (OTKA) Grant No. 922 to J.Sz., Ministry of Culture and Education of Hungary Grant No. FKFP 1384 to J.Sz., PECO Cell Cycle Network Program Grant No. CEC ERB CIPD CT 94 0049 to David M. Glover and J.Sz.