To identify early requirements for zygotic gene activity in Drosophila, we used compound autosomes and autosome-Y translocations to generate embryos deficient for cytologically defined portions of the genome. No obvious gross morphological defects were observed in any deficiency class until the beginning of cycle 14. Only seven autosomal regions were identified with discrete effects visible prior to the onset of gastrulation. These regions include genes with locus-specific effects on the clearing of the cortical cytoplasm during early cycle 14, (22AB), the initiation of the slow and fast phases of cellularization (26BF and 40AC, respectively), the apical-basal distribution of nuclei during cycle 14 (71C–75C) and the closing off of furrow canals during cellularization (100AC). The distal tip of the third chromosome also contains two loci (99DF and 100AC) whose deletion causes multiple nuclei to be cellularized into single cells, a phenotype similar to that produced in embryos totally lacking the X-chromosome.

The cellular transformations that occur in most early embryos are rapid and synchronous, and they occur on a scale not duplicated later in development. Embryos from a variety of different species have therefore been used to study basic cellular phenomena like DNA replication, cytokinesis and cytoplasmic localization. Moreover, since the genotype of an individual is established only at fertilization, the very early stages of embryogenesis provide a useful system for identifying specific gene activities required for these events.

In Drosophila, most of the gene products required for cleavage and blastoderm formation are already present as RNA or proteins in the unfertilized egg (Zalokar & Erk, 1976; Arking & Parente, 1980; Gutzeit, 1980). The egg cell can thus be likened to a complex machine poised for a series of morphological transformations that must occur at precise times and often in precise regions of the egg. What triggers these events or causes them to occur at particular times is not known. One obvious candidate would be the controlled synthesis of zygotic gene products at particular stages of development. Significant amounts of RNA begin to be made in Drosophila during the syncytial blastoderm stage and the levels of newly synthesized RNA are substantial by cellularization (Zalokar, 1976; McKnight & Miller, 1976; Anderson & Lengyel, 1981; Edgar & Schubiger, 1986). Few such early-acting genes have been identified genetically, however, and very little is known about their function.

In a preceding paper (Wieschaus & Sweeton, 1988), we have studied the role of X-chromosomal gene activity in early development and have identified a single locus at 6F1 –2 that is required for the establishment of the hexagonal actin arrays associated with cytokinesis in early cycle 14. To identify other loci involved in the cellularization process, we have examined embryos that lack a substantial portion of each autosome, but are derived from mothers that are genetically normal. In Drosophila, such embryos can be generated using compound chromosomes and translocations (Rasmussen, 1960; Scriba, 1967, 1969). In the following paper, we use such chromosomes to create overlapping deletions for the entire Drosophila genome. We show that even very large deletions develop normally to the late syncytical stages and only begin to show discrete abnormalities during cycle 14. We identify seven early-acting regions responsible for these abnormalities and analyse the phenotypes the corresponding deletions produce during cycle-14 cellularization.

Genotypes and stocks used

Two attached autosome stocks were used to generate embryos deficient for entire chromosome arms: RM(2L);RM(2R), marked with vermillion (= C(2)v, Bowling Green Stock Center no. nl6); and C(3L)st;C(3R)e, a compound third chromosome obtained from the University of Indiana, Bloomington. The meiotic segregation pattern of each compound chromosome was determined by mating males or females from the compound stock to flies heterozygous for four different embryonic lethals (Nüsslein-Volhard & Wieschaus (1980); i.e. odd-skipped · evenskipped/paired ·Kruppel in the case of the second chromosome, or hairy hedgehog/hunchback knirps in the case of the third chromosome). Aneuploid embryos from such crosses die but make cuticle, which can be scored for segmentation phenotypes. Embryos with normally segmented cuticle will arise only when the parent from the compound stock contributes both halves of the compound to the gamete. The fraction of the total embryos derived from the cross with normal segmentation thus provides a measure for the frequency with which the right and left arms of the compound segregate to the same pole during meiosis (see Fig. 1).

Fig. 1.

Generation of embryos deficient for large regions of the second chromosome. (A) Compound autosome stocks. Instead of the normal diploid arrangement in which the left and right arms of a given chromosome are attached to the same centromere, compound autosomes have both left arms attached to one centromere and both right arms attached to the other (Rasmussen, 1960). These individuals contain the normal genetic complement; their gametes, however, contain either two left or two right arms of the attached autosome, or all four arms or none. Segregational analysis using embryonic lethals (Materials and methods) suggest that virtually all the female gametes on the compound-2 stock are either 2R+2L+ or 2R”2L+, but that 44% of the male gametes are 2R+2L+ or 2R”2L-. Similar values are obtained for the compound-3 stock (95% and 48%, respectively). In either stock, inbreeding produces about 50% normal embryos, 25 % with no right arm of the attached autosome and 25 % with no left arm of that autosome. The latter two classes of embryos will show defects as soon as a zygotic gene product from the missing arm is required for normal development. (B) Crosses between compound autosome stocks and 2Y translocations. To obtain smaller deletions, compound autosome females were crossed to males bearing Y-autosome translocations with breakpoints at various points along the chromosome. As long as the required region is distal to the translocation breakpoint, one eighth of the embryos from the cross will lack the relevant gene and exhibit the abnormal phenotype; a cross in which no embryos show the abnormality indicates that the gene is proximal to this breakpoint.

Fig. 1.

Generation of embryos deficient for large regions of the second chromosome. (A) Compound autosome stocks. Instead of the normal diploid arrangement in which the left and right arms of a given chromosome are attached to the same centromere, compound autosomes have both left arms attached to one centromere and both right arms attached to the other (Rasmussen, 1960). These individuals contain the normal genetic complement; their gametes, however, contain either two left or two right arms of the attached autosome, or all four arms or none. Segregational analysis using embryonic lethals (Materials and methods) suggest that virtually all the female gametes on the compound-2 stock are either 2R+2L+ or 2R”2L+, but that 44% of the male gametes are 2R+2L+ or 2R”2L-. Similar values are obtained for the compound-3 stock (95% and 48%, respectively). In either stock, inbreeding produces about 50% normal embryos, 25 % with no right arm of the attached autosome and 25 % with no left arm of that autosome. The latter two classes of embryos will show defects as soon as a zygotic gene product from the missing arm is required for normal development. (B) Crosses between compound autosome stocks and 2Y translocations. To obtain smaller deletions, compound autosome females were crossed to males bearing Y-autosome translocations with breakpoints at various points along the chromosome. As long as the required region is distal to the translocation breakpoint, one eighth of the embryos from the cross will lack the relevant gene and exhibit the abnormal phenotype; a cross in which no embryos show the abnormality indicates that the gene is proximal to this breakpoint.

The autosome-Y translocations that were used to produce embryos with smaller genomic deficiencies are listed in Table 1. When necessary, the segregation pattern of individual translocations was determined by mating translocation males to females heterozygous for lethal mutations that had been mapped previously to the right or left arm of the chromosome affected by the translocation (generally odd-skipped and even-skipped on the second chromosome, or hairy and hedgehog on the third chromosome). The frequency with which the zygotes displayed the segmental phenotype associated with the mutant on the given arm was used to determine the segregation pattern of the particular translocation tested.

Table 1.

Chromosomal aberrations used to generate deficiency embryos

Chromosomal aberrations used to generate deficiency embryos
Chromosomal aberrations used to generate deficiency embryos

Phenotypic characterization of living embryos

Living embryos were covered with Voltalef 3S Halocarbon oil to make the egg shell transparent and their development was followed in bright-field optics using a stereomicroscope (×50). About 100 –200 eggs were collected from a given stock or cross, observed at 10 –15 min intervals and classified according to phenotype. A fraction of the eggs from most collections showed no sign of development and appeared to be unfertilized. While it is possible that some abnormal embryos might show this appearance, the large fluctuations in the frequency of apparently unfertilized eggs in different collections (19%, 13%, 15%, 7%, 5%, 7% for the second chromosomal stocks and 9 %, 5 %, 12 % and 2% for the third chromosomal stock) can best be interpreted as due to variable fertilization in the compound stocks; abnormalities associated with a particular embryonic genotype would be expected to give more consistent percentages. When these ‘unfertilized’ eggs are not considered, all expected deficiency classes are observed and at frequencies (25–29%) expected for normal meiotic segregation. This argues that, whether or not these eggs are in fact unfertilized, they do not represent the particular deficiency class whose members are the subject of these experiments.

Photomicrographs and time-lapse video tapes of optical sections of individual embryos at x200 or X400 were made following previously published procedures (Wieschaus & Nüsslein Volhard, 1986). The conclusions presented in the results section are based on videotapes of 66 normal embryos, 9 2L embryos, 4 2R embryos, 19 3L and 18 3R embryos.

To follow the density and spacing of surface nuclei during cleavage cycles 10 through 14, we used a compound microscope and epi-illumination created by directing the light beam from a fibre optic source across the surface of an otherwise unilluminated embryo. For maximum resolution, embryos were dechorionated with dilute sodium hypochlorite, viewed with a ×lOO oil immersion objective and the image enhanced by transmission through the video camera/time-lapse recorder system.

Whole-mount preparations

The F-actin and nuclear distributions during late cleavage stages were analysed by staining fixed whole-mount embryos with rhodamine-labelled (RH) phalloidin and Hoechst 33258 dye, following previously described procedures (Wieschaus & Nüsslein Volhard, 1986). The stained embryos were examined at ×16, ×25 and ×lOO using a Zeiss compound microscope with fluorescence optics.

In most cases, embryos were stained with both Hoechst and RH-phalloidin. However, in experiments where the frequency of mitotic abnormalities in different stocks were to be compared, embryos from both stocks were stained with Hoechst 33258 but only one of the stocks was stained with RH-phalloidin. The embryos were then mixed and transferred to glass slides. The cleavage pattern of each embryo was scored using the u.v. wave length (based on the Hoechst stain) and the stock origin of the embryo was determined using the RH-phalloidin only after the embryos had been classified.

Sectioned material for light microscopic and electron microscopic analysis was prepared following previously published protocols for fixation and embedding (Wieschaus & Nüsslein Volhard, 1986).

To identify early requirements for zygotic gene activity, we have followed the development of embryos deficient for defined portions of the Drosophila genome. Many of the most characteristic events of early Drosophila development can be scored directly in living embryos (Foe & Alberts, 1983; Wieschaus & Nüsslein-Volhard, 1986). These include (1) the emergence of the polar buds and pole cells following the eighth and ninth nuclear division; (2) the synchronous contractions of the cortical cytoplasm as the nuclei undergo their tenth through thirteenth cleavage divisions; (3) the clearing of the cytoplasm at the outer edge of the syncytial blastoderm as yolk and lipid droplets are transported basally and (4) the even descent of the membranes between regularly aligned nuclei during cellularization.

No single region on either the second or the third chromosome is required for gross normal morphology up to the onset of cycle 14

The second and third chromosome each represent about 40 % of the genome in Drosophila and judging from the number of bands on polytene chromosomes should each contain about 2000 lethal complementation groups. Both are metacentric in morphology, with similarly sized right and left arms attached at the centromere. To determine whether activity of either chromosome is required during the cleavage or cellularization, we generated deficiency embryos using attached autosomes in which the two right arms and the two left arms, normally present in a diploid, were fused to each other (Fig. 1A). When such females undergo meiosis, half the resulting gametes lack all copies of the right arm and half lack all copies of the left. Meiotic segregation in males gives rise to these gametes, but also produces sperm that have both right and left arms or neither arm (Baldwin & Chovnick, 1967; Wright, 1970). Ultimately about one quarter of the embryos derived from a cross between compound males and females will lack the right arm and one quarter will lack the left arm (Fig. 1).

Although each compound stock produced two morphologically distinct classes of deficiency embryos by gastrulation, virtually all embryos from the stocks develop normally until cycle 14, at least judging from their appearance under a stereomicroscope. To analyse early development in greater detail, we made time-lapse video tapes of developing embryos from compound stocks at stages between pole cell formation and gastrulation. Embryos were videotaped using either optical cross-sections at ×40 magnification and epi-illumination to follow the dividing nuclei in the surface cortex (Fig. 2, see also Materials and methods). Based on their subsequent phenotypes at gastrulation (see below), it was possible to distinguish the deficiency embryos from their heterozygous siblings. The data in Table 2 show that wildtype and deficiency embryos have cleavage cycles of similar length and duration. The density of the nuclei in deficiency embryos at interphase of each cycle is basically normal, although some deviation from a uniform spacing was observed in 3L embryos (and possibly 2R, see below). Deficiency embryos also begin the major morphogenetic movements of gastrulation at approximately the same times as wild-type embryos. Taken together, these results identify no region on either the second or the third chromosome required for gross normal morphology prior to cycle 14. They argue in turn that the major morphological transformations that occur during cleavage rely almost entirely on maternal gene products already in the egg at fertilization. A similar conclusion is suggested by the observation that embryos injected with α-amanitin develop normally to the beginning of cycle 14 (Edgar et al. 1986). lite early cleavage defects described for various deficiency embryos by other authors (Scriba, 1967, 1969: reviewed in Wright, 1970) may be attributed to dominant maternal effects of the heterozygous mothers (Garcia-Bellido et al. 1983); or to the difficulty in staging abnormal embryos that have been fixed and sectioned (see Wieschaus & Sweeton, 1988).

Table 2.

Length of cleavage cycles and density of surface nuclei between cycles 10 and 14 of embryos derived from the compound-2 and compound-3 stocks

Length of cleavage cycles and density of surface nuclei between cycles 10 and 14 of embryos derived from the compound-2 and compound-3 stocks
Length of cleavage cycles and density of surface nuclei between cycles 10 and 14 of embryos derived from the compound-2 and compound-3 stocks
Fig. 2.

Epi-illumination photomicrographs of surface nuclei during cleavage cycle 10 –14. Normal embryo (left panel); 2R embryo (middle panel); 3L embryo (right panel). Note the irregular spacing and reduced number of nuclei in the 3L embryo, compared to the more regular distributions in normal and 2R embryos.

Fig. 2.

Epi-illumination photomicrographs of surface nuclei during cleavage cycle 10 –14. Normal embryo (left panel); 2R embryo (middle panel); 3L embryo (right panel). Note the irregular spacing and reduced number of nuclei in the 3L embryo, compared to the more regular distributions in normal and 2R embryos.

Embryos deficient for the left arm of the second chromosome are defective in transport of lipid droplets from the periphery

28% of the embryos from the compound-2 stock showed a three-layered (‘halo’) appearance early in cycle 14 (Fig. 3A,C), due to an altered distribution of the cortical cytoplasm prior to cellularization. In wildtype embryos, the depth of the cortical cytoplasm increases after the thirteenth cleavage as lipid and yolk are transported basally (Fullilove & Jacobson, 1971; Lundquist & Emanuelsson, 1979; Foe & Alberts, 1983). In ‘halo’ embryos from the com-pound-2 stock, yolk transport occurs normally but the lipid droplets remain in the peripheral cytoplasm immediately below the nuclei (Fig. 4A,B). The refraction of these lipid droplets darkens the cytoplasm when embryos are viewed with transillumination, causing a halo of shaded cytoplasm below the nuclei. Below the lipid halo, the cytoplasm next to the yolk is poor in lipid, producing a second clear layer immediately adjacent to the yolk. The phenotype becomes unambiguous in time-lapse videos 10 min after the beginning of cycle 14, but since the cytoplasmic clearing begins much earlier, it is likely that a quantitatively less striking phenotype may be detectable in early cycles.

Fig. 3.

Living embryos deficient for large regions of the second chromosome. (A) Normal embryo during cellularization and early gastrulation; (B) 2Rembryo during late cellularization and early gastrulation. Note absence of the ventral furrow and presence of normal posterior midgut; (C) 2L” embryos during midcycle 14 and gastrulation. Note obvious ‘halo’ of lipid-rich cytoplasm and abnormal attempt at posterior midgut formation and gastrulation (—>); (D) Embryo deficient for the tip of the second chromosome to 26A, during early and late cellularization. Note ‘halo’ and uniform cellularization ring; (E) Embryo deficient for the tip of the second chromosome to 39C, during late cellularization. Note ‘halo’ and irregular cellularization ring; (F) Embryo homozygous for Df(2)ast-2 (=21Dl-2; 22B2-3). Note ‘halo’.

Fig. 3.

Living embryos deficient for large regions of the second chromosome. (A) Normal embryo during cellularization and early gastrulation; (B) 2Rembryo during late cellularization and early gastrulation. Note absence of the ventral furrow and presence of normal posterior midgut; (C) 2L” embryos during midcycle 14 and gastrulation. Note obvious ‘halo’ of lipid-rich cytoplasm and abnormal attempt at posterior midgut formation and gastrulation (—>); (D) Embryo deficient for the tip of the second chromosome to 26A, during early and late cellularization. Note ‘halo’ and uniform cellularization ring; (E) Embryo deficient for the tip of the second chromosome to 39C, during late cellularization. Note ‘halo’ and irregular cellularization ring; (F) Embryo homozygous for Df(2)ast-2 (=21Dl-2; 22B2-3). Note ‘halo’.

Fig. 4.

Electron micrographs of lipid clearing during early cycle 14. (A) Normal embryo; (B) deficiency 2Lembryo (segregant from the C(2)v stock) and (C) Df(2)ast-2 homozygous embryo.

Fig. 4.

Electron micrographs of lipid clearing during early cycle 14. (A) Normal embryo; (B) deficiency 2Lembryo (segregant from the C(2)v stock) and (C) Df(2)ast-2 homozygous embryo.

The zygotically active gene whose absence causes halo formation was localized on the left arm of the second chromosome by crossing C(2)v females with T(Y;2) translocation males with breakpoints in that arm. One eighth of the embryos from such crosses will be deficient for all genes distal to the translocation breakpoint (Fig. 1B). Halo embryos were produced by males with translocation breakpoints from 39C to 22A, but not by males with breakpoints in 21E (Table 1). These results suggest that the locus responsible for the halo phenotype lies between 21E and 22A. To localize further the gene, embryos were collected from stocks heterozygous for deficiencies in the 21–23 region (Table 1). The definitive halo phenotype (Fig. 3F, Fig. 4C) was observed only in embryos from the osf-l (21C7–8,23Al–2) and ast-2 (21Dl–2,22B2–3) stocks, where they represented one quarter of the total embryos from the stock. They were not observed in embryos homozygous for other deficiencies that substantially overlap the distal regions of ast-1 and ast-2 (Table 1), localizing the gene to the 22A6–B2 region.

In spite of their abnormal lipid transport, ast-2 homozygotes cellularize and gastrulate normally. At 24 h, mouth parts and segmentation are visible, but none of these embryos hatch. These later phenotypes can be attributed to other genes within the ast-2 deficiency since they are also produced by slightly overlapping deficiencies which do not lack bands in 22A6–B2 responsible for the ‘halo’ phenotype.

Genes affecting cellularization on the second chromosome

Cellularization in Drosophila normally begins immediately after the nuclei enter interphase of cycle 14 (Mahowald, 1963; Fullilove & Jacobson, 1976; Turner & Mahowald, 1976; Lundquist & Emanuelsson, 1979) and is completed in about 55 min at room temperature. It occurs in two phases: an initial slow phase thought to proceed solely by addition of new membrane (Turner & Mahowald, 1976; Warn & MacGrath, 1983), followed by a fast contractile phase in which membrane from the surface microvilli appears to be pulled down into the cleavage furrows.

The 2L halo embryos derived by inbreeding C(2)v males and females show normal nuclear and actin localization through early stage 14, but they do not form cells. Since Df(2)ast-2 homozygotes cellularize and gastrulate normally, other regions on 2L must be responsible for the abnormalities observed in embryos that lack the entire arm. By crossing C(2)v females to various translocation males (Table 1), the cellularization defects observed in 2L” embryos could be localized to two additional zygotically active regions, one between 26B and 26F and the other between 39C and the centromere.

The 2L-deficient embryos from translocation crosses with breakpoints at or distal to 26B form normal cellular blastoderms (Fig. 3D); embryos lacking regions distal to 26F only begin to cellularize about 40min into the cycle (Fig. 3E; Fig. 5). The resulting cellularization is not uniform over the entire surface of the embryo, possibly because the embryo begins to gastrulate only about 15 min after the visible onset of cellularization, at a time when the membranes have reached less than half their normal extent. The staging and rapidity with which the membranes advance suggest that the abnormal cellularization in these embryos corresponds to ‘fast’, contractile, phase. The absence of cellularization during the first 40 min of cycle 14 indicates the existence of a gene in 26B-F which is required for the initial slow phase involving membrane addition.

Fig. 5.

Cellularization rates in normal embryos and embryos deficient for either the proximal or distal regions of 2R. (A) Embryos deficient for the distal portion of the left arm of the second chromosome (=○), obtained by mating compound-2 females to either T(Y;2)L138,39C or T(Y;2)A80,35A males. Two of the three embryos recorded were selected based on their halos phenotype. Since the first 15 min of cycle 14 were not videotaped, the data from these embryos have been aligned with the third embryo using the point when they began gastrulation. The data for normal embryos (=•) were derived from six sibling embryos from the same cross. (B) Embryos deficient for the proximal regions of the left arm of the second chromosome (=○), obtained by mating compound-t2 females to T(Y;2)G,26B6-C1 males. Three embryos were measured. Values for normal embryos (=•) are based on five sibling embryos from the same cross. The depth of the cellularization furrow was measured on the dorsal side where it is not distorted by ventral furrow formation. Error bars indicate one standard deviation.

Fig. 5.

Cellularization rates in normal embryos and embryos deficient for either the proximal or distal regions of 2R. (A) Embryos deficient for the distal portion of the left arm of the second chromosome (=○), obtained by mating compound-2 females to either T(Y;2)L138,39C or T(Y;2)A80,35A males. Two of the three embryos recorded were selected based on their halos phenotype. Since the first 15 min of cycle 14 were not videotaped, the data from these embryos have been aligned with the third embryo using the point when they began gastrulation. The data for normal embryos (=•) were derived from six sibling embryos from the same cross. (B) Embryos deficient for the proximal regions of the left arm of the second chromosome (=○), obtained by mating compound-t2 females to T(Y;2)G,26B6-C1 males. Three embryos were measured. Values for normal embryos (=•) are based on five sibling embryos from the same cross. The depth of the cellularization furrow was measured on the dorsal side where it is not distorted by ventral furrow formation. Error bars indicate one standard deviation.

The gene in 26B-F does not, however, account for the entire nullo-2L phenotype, since no cellularization at all is observed in nullo-2L embryos and halo embryos derived from translocation males with breakpoints in the centromeric heterochromatin. Since translocation males with breakpoints in 39C produce the same cellularization phenotype as those with breakpoints in 26F, the gene required for the late cellularization occurring in both genotypes must lie between 39C and the centromere. To obtain embryos deficient for the proximal regions of 2L, we mated compound-2 females to males carrying an insertional translocation T(Y;2)G (36B6-C; 40F). The most obvious abnormality in embryos deficient for the proximal 2L is the shallowness of the blastoderm cell layer as the animal begins to gastrulate. Time-lapse videos demonstrate that this is largely due to the failure of the invaginating membranes to accelerate their rate of elongation during the last 15 min of cellularization (Fig. 5B). The phenotype is consistent with a role of the proximal gene in the fast contractile phase of cellularization, but judging from the rates of cellularization depicted in Fig. 5B, the underlying defect may extend into the earlier phase of cellularization as well.

Embryos deficient for 2R cellularize normally but fail to make ventral furrows at gastrulation

The compound-2 stock produces a second class of abnormal embryos that fail to form a ventral furrow at gastrulation (Fig. 3D). Such embryos begin a posterior midgut invagination normally and form cephalic furrows, but those structures also soon become abnormal. Identical phenotypes are produced in one eighth of the embryos when C(2)v females are mated to translocation males with breakpoints in proximal 2R, confirming that this abnormal class results from deficiencies of the right arm of the second chromosome.

The normal appearance of 2R embryos through cellularization was unexpected, given the observation by Karr et al. (1985) that embryos homozygous for a mutation on 2R (engrailed) become abnormal during cleavage divisions. Using our mounting procedures, we could not detect the reported right-left asymmetry’ of the pole cells in living embryos shown subsequently deficient for 2R, nor have we observed any other consistent pregastrulation defects in the embryos we videotaped (Fig. 2, Table 2). To detect more subtle defects, we compared the frequency of mitotic irregularities in fixed embryos from the C(2)v stock with those found in wild-type stocks. Because the phenotype reported for engrailed overlaps patterns we have seen in wild-type stocks (Fig. 6), we used procedures that allowed us to score individual embryos without knowing their genotype or stock of origin (Materials and methods). When summed over cycles 10–14 (Table 3), the C(2)v stock embryos showed approximately the same overall percentages of asynchronous divisions and irregular patterns as wild type. There was, however, a significant increase of abnormal patterns in embryos from the C(2)v stock fixed during cycles 12 and 13 (Table 3). The increased frequency of abnormalities during these cycles is compatible with their being a deficiency phenotype, if one assumes that only one of the two second chromosomal arms affects mitotic synchrony. Overall the effect is rather subtle and we have not used translocations to localize it further. In the context of the present paper, it does indicate a lower limit in the ability of our procedures to detect early defects.

Embryos deficient for 3L show an abnormal distribution of nuclei to the central cytoplasm

Approximately a quarter of the embryos from the compound-3 stock show a ‘fuzzy’ appearance at the interface between the yolk and cortical cytoplasm in the earliest stages of cycle 14 (Fig. 7). In videotapes made of the posterior pole at ×40, the abnormality is apparent as soon as the nuclei become again visible after the 13th division. When viewed using epiillumination, the spacing of the nuclei in the surface is irregular and not all nuclei can be resolved in the same focal plane. In sectioned embryos at this stage, many nuclei are observed below the surface, apparently being transported to the central yolky core (Fig. 8A,B). A large fraction (85%) of the ‘fuzzy’ embryos undergo an additional round of division midway through the cellularization of cycle 14 (Fig. 7, middle panel, second frame), perhaps in response to the reduced nuclear density at the surface. After the extra division is completed, the embryos continue cellularization and attempt to make a ventral furrow and a posterior midgut, but the basic appearance of a ‘fuzzy’ yolk mass remains.

Table 3.

Cleavage patterns in embryos derived from the compound-2 stock and thus potentially deficient for either the right or left arm of the second chromosome

Cleavage patterns in embryos derived from the compound-2 stock and thus potentially deficient for either the right or left arm of the second chromosome
Cleavage patterns in embryos derived from the compound-2 stock and thus potentially deficient for either the right or left arm of the second chromosome
Fig. 6.

Abnormal cleavage patterns observed in embryos from wild-type and compound-2 stocks. Wave-like mitotic patterns observed in Oregon-R stock (A) and in compound stock (B). Asynchronous mitotic pattern observed in Oregon-R embryo (C) and in compound stock (D). Variable irregular spacing of nuclei in Oregon R stock (E) and in compound stock (F,G). Dense internalized nuclei at posterior pole of Oregon R embryo (H) and in compound stock (I)-

Fig. 6.

Abnormal cleavage patterns observed in embryos from wild-type and compound-2 stocks. Wave-like mitotic patterns observed in Oregon-R stock (A) and in compound stock (B). Asynchronous mitotic pattern observed in Oregon-R embryo (C) and in compound stock (D). Variable irregular spacing of nuclei in Oregon R stock (E) and in compound stock (F,G). Dense internalized nuclei at posterior pole of Oregon R embryo (H) and in compound stock (I)-

Fig. 7.

Development of cycle-14 embryos deficient for various portions of the third chromosome. Posterior third of embryo, photograph taken directly from the time-lapse videomonitor during playback. Normal embryo (left panel); 3L embryo (middle panel); 3R (right panel).

Fig. 7.

Development of cycle-14 embryos deficient for various portions of the third chromosome. Posterior third of embryo, photograph taken directly from the time-lapse videomonitor during playback. Normal embryo (left panel); 3L embryo (middle panel); 3R (right panel).

Fig. 8.

Distribution of nuclei and cortical actin during cycle 14 in embryos deficient for either the right or the left arm of the third chromosome. Sections through surface cortex of normal (A), 3L (B) and 3R” (C) embryos, showing the transport of nuclei from the surface of 3L” embryos, and the entrapment of nuclei at the cellularization furrow in 3R” embryos. Surface views of 3L embryo (upper) and 3R embryo (lower) stained to show nuclei (D) and actin (E) distributions. Note the abnormal clustering of nuclei and irregular actin arrays in the 3R embryo, and the less-densely spaced nuclei, but more regular actin arrays in the 3L embryo. In optical cross-sections (F), the nuclei of the 3R” embryo appear trapped at the cell interface, whereas many of the nuclei in the 3L embryo have been displaced to the cytoplasm-yolk interface. Embryos hemizygous for the DF(3)tlF show the same pinched nuclear morphology (G) at the cellularization front (H) observed in 3R- embryos. Surface views at (I) and immediately above (J) the cellularization front show the small pinched off openings at the front. Note the occasional larger openings when more than one nucleus is cellularized into the same cell. Abnormal distributions of actin (K) and clustered nuclei (L) in embryos homozygous for Df(3R)X3F.

Fig. 8.

Distribution of nuclei and cortical actin during cycle 14 in embryos deficient for either the right or the left arm of the third chromosome. Sections through surface cortex of normal (A), 3L (B) and 3R” (C) embryos, showing the transport of nuclei from the surface of 3L” embryos, and the entrapment of nuclei at the cellularization furrow in 3R” embryos. Surface views of 3L embryo (upper) and 3R embryo (lower) stained to show nuclei (D) and actin (E) distributions. Note the abnormal clustering of nuclei and irregular actin arrays in the 3R embryo, and the less-densely spaced nuclei, but more regular actin arrays in the 3L embryo. In optical cross-sections (F), the nuclei of the 3R” embryo appear trapped at the cell interface, whereas many of the nuclei in the 3L embryo have been displaced to the cytoplasm-yolk interface. Embryos hemizygous for the DF(3)tlF show the same pinched nuclear morphology (G) at the cellularization front (H) observed in 3R- embryos. Surface views at (I) and immediately above (J) the cellularization front show the small pinched off openings at the front. Note the occasional larger openings when more than one nucleus is cellularized into the same cell. Abnormal distributions of actin (K) and clustered nuclei (L) in embryos homozygous for Df(3R)X3F.

When compound-3 females were crossed with translocation males with breakpoints from 75C or 78CD, one eighth of the progeny showed the ‘fuzzy’ phenotype, localizing the required gene to the left arm of the third chromosome (Table 1). Under the dissecting microscope, the embryos derived from crosses using more distal breakpoints (71C. 65D) were phenotypically normal until the onset of gastrulation. This argues against the existence of any loci with early zygotic effect on gross morphology distal to 71C and suggests that the ‘fuzzy’ locus must lie somewhere between 71C and 75CD.

Development of embryos deficient for 3R

A second class of embryos from the compound-3 stock becomes distinct only as cell membranes descend between nuclei about 15 min into cycle 14. At this point, the nuclei in mutant embryos become disorganized and some are shifted out of the uniform single layer (Fig. 7C). In sectioned material, these nuclei appear constricted in the middle (Fig. 8C,G), as though they are being carried by the advancing cellularization front into the interior of the embryo. When the embryo has reached mid-cellularization stages, two layers of nuclei can be observed, the more basal layer coinciding with the advancing front of cellularization (Fig. 8E). When cell membranes have reached their normal depth, the embryos begin an abnormal gastrulation and soon assume a characteristic morphology due to their failure to form a posterior midgut (Fig. 7, right panel, last frame).

Normally during cellularization, the cleavage furrows only begin to widen and pinch off cells as they approach the prospective yolk sack. In 3R- embryos, this pinching-off occurs early during the cellularization process, trapping nuclei and dragging them from the surface (Fig. 8G-J). All of the crosses between C(3)se females and T(Y;3R) males produced embryos that showed the double-layered nuclear phenotype, a result that localized the gene responsible beyond the most distal of the translocation breakpoints we used (100A5 –6). By examining the development of embryos homozygous for various deficiencies in that region, we found that the nuclear layering was produced in embryos homozygous for Df(3)tlle (100A1 –2; 100C1), but not in Df(3)tlF (99F1 –2; 100B4 –5), putting the locus in the six bands between 100B4 and 100C1. A cellularization defect mapping to Df(3)tlle but not Df(3)tllg has been described previously (Mahoney & Lengyel, 1987). 3R” embryos show a second phenotype very similar to that observed in nullo-X embryos (Wieschaus & Sweeten, 1988). Early during cellularization multiple nuclei are often enclosed within the same cell membranes. Subsequently, the ventral side of 3R- embryos becomes particularly irregular and, like nullo-X embryos, the yolk sac begins ‘oozing’ out to the surface at the posterior ventral pole (Fig. 7). Rh-phalloidin-stained whole mounts of early cycle-14 embryos show the same irregular patterns of actin and nuclei observed in nullo-X embryos, arguing that the abnormal cellularization may arise by similar disruptions in the hexagonal actin arrays responsible for the initial positioning of the cell membranes.

This phenotype is produced in all crosses using translocations with breakpoints proximal to 98F (Table 1) and in somewhat diminished form by even the most distal translocations used in this study. Analysis of deficiencies for the distal region of 3R identify at least two regions required for that aspect of normal cellularization (JDf(3)XF3, 99D1 –2 to 99E1, Fig. 8K,L; and Df(3)tlF, 100A1 –2 to 100C, Fig. 8G-J). The 6F1 –2 region of the X-chromosome and the distal tip of the third chromosome seem to be the only regions of the Drosophila genome required for this aspect of cellularization since embryos deficient for 2L (21A –39C), 2R (41 –60F) and 3L (61 –80) show normal actin arrays at the beginning of cycle 14 (data not shown).

By generating synthetic deficiencies for cytologically defined portions of the Drosophila genome, we have examined the early requirements for zygotic gene activity during early embryogenesis. The largest deficiencies represented a single chromosomal arm, about 20% of the entire genome. All such deficiencies allowed development to the beginning of cycle 14, arguing that the first thirteen cleavage cycles run predominantly, if not exclusively, using gene products supplied by the mother during oogenesis. We could reliably detect early defects only during the initial stages of cycle 14 and several fairly large deletions (e.g. 41A to 60F; 61A to 71CD) allowed normal development until the onset of gastrulation. Even in those cases where the deficient embryos became abnormal, the defects were initially fairly discrete, an observation that can be most easily explained if they represent the requirement for only a single gene. In most cases, the translocation mapping allowed localizing the required gene to small regions of the chromosome, often only a few bands in length.

Overall our data argue for the existence of seven autosomal genes whose activities are required prior to the onset of gastrulation. The reliability of that number obviously depends on the resolution of our morphological techniques. Regions causing more subtle effects (e.g. those reported for engrailed, Karr et al. 1985) would have been missed. Moreover, because of the manner in which smaller deletions were generated, more proximal regions on a given arm could usually only be examined in embryos also deficient for all distally located genes. Given the cluster of early-acting genes on the tip of 3R, for example, more proximal loci might be masked by the cellularization phenotypes associated with the distal genes. On the other hand, embryos totally deficient for 3R from the compound stock do not seem significantly more abnormal during cellularization than embryos deficient from 98F to the tip.

Although the primary goal of the present study was to identify requirements for zygotic gene activity during cellularization, the relative scarcity of such early-acting loci allowed us to examine many of the large deficiency class embryos for defects during gastrulation. In general, we have not identified any new loci in this way; i.e. most of the abnormalities we observed during gastrulation could be attributed to genes previously identified in our EMS screens (Nüsslein-Volhard et al. 1984; Jürgens et al. 1984; Wieschaus et al. 1984). For example, using various translocations (Table 1), we have shown that the failure of nullo-2R embryos to make a ventral furrow is due to a gene between 59A and 59F, most likely the twist locus (Nüsslein-Volhard et al. 1984). Other gastrulation effects in the compound-2 stock could be similarly attributed to the decapentaplegic (22D) and the snail (35CD) loci. The only case where a gastrulation phenotype could not be explained using previously known point mutants is the failure of embryos deficient for distal 3R to make a posterior midgut (Fig. 7C). This might be a secondary consequence of the cellularization defects associated with that region, but it is also possible that the phenotype is due to a previously undescribed gastrulation locus. In general, however, our results confirm the scarcity of early-acting loci in Drosophila, both for cellularization and gastrulation.

Perhaps the most remarkable thing is not that the number of early loci is so small but that their phenotypes are so unique and specific for particular aspects of cellularization. The discrete nature of the phenotypes argue that the wild-type functions associated with each locus may be distinct and therefore useful in dissecting early cellularization into its component parts. An analysis of the slow and fast phases of membrane elongation is particularly promising in this respect, but general usefulness of the approach may extend to other aspects of cellularization as well.

In contrast to the mutations identified in our previous EMS mutagenesis experiments (Nüsslein-Volhard et al. 1984; Jürgens et al. 1984; Wieschaus et al. 1984), the defects described in this paper are not localized to spatially defined regions of the developing embryo. Given the low requirements for zygotic gene activity in Drosophila, it had seemed possible that the embryo might use early transcription predominantly to localize gene products to particular cells or primordia. This view was consistent with the localized defects and transcript distributions described for many previously analysed embryonic lethals. We were surprised to find zygotic requirements during cellularization, given that the process occurs uniformly over the entire surface of the embryo. A priori, it is not obvious why these gene products, like most other components of cellularization, would not be supplied maternally. One possibility is that their requirements are localized, but localized in time rather than space. The sequence of events during cellularization is complex enough that the individual phases may require specific zygotic initiators. Given the stage-specific nature of their phenotypes, the early-acting loci we have identified are good candidates for zygotic initiators.

We are grateful to Jim Kennison for supplying most of the translocation stocks used in this study. Judy Lengyel gave us the deficiencies in distal 3R and Paul Mahoney shared unpublished information with us. David Roberts and Barry Ganetzky gave us stocks for the distal and proximal regions of 2L respectively. We are also grateful for our colleagues at Princeton for comments and criticisms during the course of the experiments. These experiments were supported by NIH grant no. HD155587 to E.W.

Anderson
,
K. V.
&
Lengyel
,
J. A.
(
1981
).
Changing rates of DNA and RNA synthesis in Drosophila embryos
.
Devi Biol.
82
,
127
138
.
Arking
,
R.
&
Parente
,
A.
(
1980
).
Effects of RNA inhibitors on the development of Drosophila embryos permeabilized by a new technique
.
J. exp. Zool.
212
,
183
194
.
Baldwin
,
M.
&
Chovnick
,
A.
(
1967
).
Autosomal halftetrad analyses in Drosophila melanogaster
.
Genetics
55
,
277
293
.
Brittnacher
,
J. C.
&
Ganetsky
,
B.
(
1983
).
On the components of segregation distortion in Drosophila melanogaster. II Deletion mapping and dosage analysis of the SD locus
.
Genetics
103
,
659
673
.
Edgar
,
B. A.
,
Ktehle
,
C. P.
&
Schubiger
,
G.
(
1986
).
Cell cycle control by the nucleo-cytoplasmic ratio in early Drosophila development
.
Cell
44
,
365
372
.
Edgar
,
B. A.
&
Schubiger
,
G.
(
1986
).
Parameters controlling activation during early Drosophila development
.
Cell
44
,
871
877
.
Foe
,
V. E.
&
Alberts
,
B. M.
(
1983
).
Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis
.
J. Cell Sci.
61
,
31
70
.
Fullilove
,
S. L.
&
Jacobson
,
A. G.
(
1971
).
Nuclear elongation and cytokinesis in Drosophila montana
.
Devi Biol.
26
,
560
577
.
Garcia-Beludo
,
A.
,
Moscoso Del Prado
,
J.
&
Botas
,
J.
(
1983
).
The effect of aneuploidy on embryonic development in Drosophila melanogaster
.
Mol. gen. Genet.
192
,
253
263
.
Gutzeit
,
H. O.
(
1980
).
Expression of the zygotic genome in blastoderm stage embryos of Drosophila: analysis of a specific protein
.
Wilhelm Roux Arch devl Biol.
188
,
153
156
.
JÜrgens
,
G.
,
Wieschaus
,
E.
,
NÜsslein-volhard
,
C.
&
Kluding
,
H.
(
1984
).
Mutations affecting the pattern of the larval cuticle of Drosophila melanogaster. II. Zygotic loci on the third chromosome
.
Wilhem Roux’s Arch, devl Biol.
193
,
283
295
.
Karr
,
T. L.
,
Au
,
Z.
,
Drees
,
B.
&
Kornberg
,
T.
(
1985
).
The engrailed locus of D. melanogaster provides an essential zygotic function in precellular embryos
.
Cell
43
,
591
601
.
Lindsley
,
D. L.
&
Grell
,
E. H.
(
1968
).
Genetic Variations of Drosophila melanogaster. Carnegie Inst. Wash Publ
627
.
Lindsley
,
D. L.
,
Sandler
,
L.
,
Baker
,
B. S.
,
Carpenter
,
A. T. C.
,
Denell
,
R. E.
,
Hall
,
J. C.
,
Jacobs
,
P. C.
,
Miklos
,
G. L. G.
,
Davis
,
B. K.
,
Gethman
,
R. C.
,
Hardy
,
W. R.
,
Hessler
,
A.
,
Miller
,
S. W.
,
Nozawa
,
H.
,
Parry
,
D. M.
&
Gould-Somero
,
M.
(
1972
).
Segmental aneuploidy and the genetic gross structure of the Drosophila genome
.
Genetics
71
,
157
184
.
Lundquist
,
A.
&
Emanuelsson
,
H.
(
1979
).
Membrane production and yolk degradation in the early fly embryo (Calliphora erythrocephala Meig.): An ultrastructural analysis
.
J. Morph.
161
,
53
78
.
Mahoney
,
P. A.
&
Lengyel
,
J. A.
(
1987
).
The zygotic segmentation mutant tailless alters the blastoderm fate map of the Drosophila embryo
.
Devl Biol.
122
,
464
470
.
Mahowald
,
A. P.
(
1963
).
Electron microscopy of the formation of the cellular blastoderm in Drosophila melanogaster
.
Expl Cell Res.
32
,
457
468
.
Mcknight
,
S. L.
&
Miller
,
O. L.
, JR
(
1976
).
Ultrastructural patterns of RNA synthesis during early embryogenesis of Drosophila melanogaster
.
Cell
8
,
305
319
.
NÜsslein-volhard
,
C.
&
Wieschaus
,
E.
(
1980
).
Mutations affecting segment number and polarity in Drosophila
.
Nature, Land.
287
,
795
801
.
NÜsslein-volhard
,
C.
,
Wieschaus
,
E.
&
Kluding
,
H.
(
1984
).
Mutations affecting the pattern of the larval cuticle of Drosophila melanogaster. II. Zygotic loci on the second chromosome
.
Wilhelm Roux’s Arch, devl Biol.
193
,
267
282
.
Rasmussen
,
I.
(
1960
).
Attached 2R;2L0
.
Dros. Information Serv.
34
,
53
.
Roberts
,
D. B.
,
Brock
,
H. W.
,
Rudden
,
N. C.
&
Evans-Roberts
,
S.
(
1985
).
A genetic and cytogenetic analysis of the region surrounding the LSP-1 B-gene in Drosophila melanogaster
.
Genetics
109
,
145
156
.
Scriba
,
M. E. L.
(
1967
).
Embryonale Entwicklungsstorungen bei Defizienz and Tetraploidie des 2. Chromosoms von Drosophila melanogaster
.
Wilhelm Roux’ Arch. EntwMech Org.
159
,
314
345
.
Scriba
,
M. E. L.
(
1969
).
Embryonale Entwicklungsstorungen bei Nullosomie und Tetrasomie des 3. Chromosoms von Drosophila melanogaster
.
Devi Biol.
19
,
160
177
.
Strecker
,
T.
,
Merriam
,
J.
&
Lengyel
,
J.
(
1988
).
Graded requirements for the zygotic gene tailless, in the brain and tail region of the Drosophila embryo
.
Development
102
,
721
734
.
Turner
,
R. F.
&
Mahowald
,
A. P.
(
1976
).
Scanning electron microscopy of Drosophila embryogenesis. I. The structure of the egg envelopes and the formation of the cellular blastoderm
.
Devi Biol.
50
,
95
108
.
Warn
,
R. M.
&
Macgrath
,
R.
(
1983
).
F-actin distribution during the cellularization of the Drosophila embryo visualized with FL-phalloidin
.
Expl Cell Res.
143
,
103
114
.
Wieschaus
,
E.
&
NÜsslein-volhard
,
C.
(
1986
).
Looking at embryos. In Drosophila, A Practical Approach
(ed.
D. B.
Roberts
).
Oxford, England
:
IRL Press
.
Wieschaus
,
E.
,
NÜsslein-volhard
,
C.
&
JÜrgens
,
G.
(
1984
).
Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. III. Zygotic loci on the X-chromosome and fourth chromosome
.
Wilhelm Roux’s Arch, devl Biol.
193
,
296
307
.
Wieschaus
,
E.
&
Sweeton
,
D.
(
1988
).
Requirements for X-linked zygotic gene activity during cellularization of early Drosophila embryos
.
Development
104
,
483
493
.
Wright
,
T. R. F.
(
1970
).
The genetics of embryogenesis in Drosophila
.
Adv. Genetics
15
,
261
395
.
Wright
,
T. R. F.
,
Hodgetts
,
B.
&
Sherald
,
A. F.
(
1976
).
The genetic of Dopadecarboxylase in Drosophila melanogaster. I. Isolation and characterization of deficiencies of the structural locus and the alpha-methyl-dopa hypersensitive locus
.
Genetics
84
,
267
285
.
Zalokar
,
M.
(
1976
).
Autoradiographic study of protein and RNA formation during early development of Drosophila eggs
.
Devi Biol.
49
,
425
437
.
Zalokar
,
M.
&
Erk
,
I.
(
1976
).
Division and migration of nuclei during early embryogenesis of Drosophila melanogaster
.
J. Microsc. Biol. Cell.
25
,
97
106
.