We have examined the early pattern of sensory mother cells in embryos mutant for six different neurogenic loci. Our results show that the neurogenic loci are required to restrict the number of competent cells that will become sensory mother cells, but are not involved in controlling the localization or the position-dependent specification of competent cells. We conclude that these loci are involved in setting up a system of mutual inhibition, which transforms graded differences within the proneural clusters into an all- or-none difference between one cell, which becomes the sense organ progenitor cell, and the other cells, which remain epidermal.

At a time when the eggs of Diptera were ‘always cited as the classical example of completely determined (mosaic) type of development, [where] the nucleus has nothing whatsoever to do with normal development except for the determination of special or superficial characters’ (Poulson, 1940), Poulson described a small deficiency of the Drosophila genome, Notch8, which has the dramatic effect of converting most of the ectoderm into a solid mass of neuroblasts. A systematic search for mutations altering neurogenesis defined six additional loci that display mutant phenotypes similar to that of Notch (N) embryos: Delta (DI), Enhancer of split (E(spl)), neuralized (neu), mastermind (mam), almondex (amx) and bigbrain (bib) (Lehmann et al. 1983). Embryos deleted for any of these loci form a large excess of neuroblasts at the expense of the ventral ectoderm. In most cases, the phenotype is aggravated if the egg derives from maternal germ cells that were homozygous for the mutation, indicating that a certain amount of maternal product is present in the oocyte (Jiménez and Campos-Ortega, 1982). The development of the peripheral nervous system from the lateral and dorsal ectoderm is also affected in neurogenic mutants: in addition to the ventral accumulation of neuroblasts, a null mutant in any one of the seven loci presents a large excess of poorly differentiated peripheral neurons and sense cells and a corresponding reduction of the lateral and dorsal epidermis to a small dorsal piece (Harten-stein and Campos-Ortega, 1986).

At least two of the neurogenic loci (N and DI) code for transmembrane proteins with EGF repeats in their extracellular domains (Wharton et al. 1985; Kidd et al. 1986; Vässin et al. 1987; Kopczynski et al. 1988), while a third one (E(spl)) would code for several functions involved in the transduction of a signal to the nucleus (Knüst et al. 1987; Hartley et al. 1988; Klambt et al. 1989). This indicates that cellular interactions might be involved in defining the number of ectodermal cells that become involved in neurogenesis (rev. in Campos-Ortega, 1988).

The formation of a sense organ in Drosophila is a progressive process (rev. in Ghysen and Dambly-Chaudière, 1989), and therefore there are many steps at which cell interactions could interfere with the number or arrangement of progenitor cells. The process begins with the acquisition of competence (proneural state) by a cluster of ectodermal cells. Competence depends on the expression of one or more genes of the proneural set, which includes daughterless and the genes of the achaete-scute complex (AS-C). One cell of the proneural cluster will emerge as a sensory mother cell (SMC), while the other cells will adopt an epidermal fate. The SMC divides to generate four or more different cells, each of which will differentiate a particular component of the sense organ. Cell interactions may help in defining the location of the initial clusters of competent cells, or may play a role in the singling out of the SMC within each proneural cluster, or may act after the emergence of the SMC, for example by controlling the proper pattern of mitosis. At which step do the different neurogenic loci act?

Most of our knowledge of neurogenic mutants is based on the analysis of late phenotypes, after cell differentiation has occurred, making it difficult to infer which step is affected by the mutations. One way to assess more directly the effect of neurogenic mutations, in the case of the central nervous system, has been to follow the pattern of expression of the proneural genes achaete and lethal of scute, both of which belong to the achaete-scute complex (AS-C). Both genes are expressed in clusters of cells at the time when neuroblasts are about to be formed. In normal flies, high levels of AS-C expression are observed in the segregating neuroblasts (Cabrera et al. 1987), while in N and DI mutants, high levels of expression are observed in all cells of the cluster (Brand and Campos-Ortega, 1989; Cabrera, 1990). These results suggest that the neurogenic loci are not involved in defining the position of the proneural clusters, but rather in the mechanism which restricts the neuroblast fate to a single cell of the cluster. This conclusion is only tentative, however, inasmuch as the presence of maternal product might have masked an earlier effect of the neurogenic mutations on the initial distribution of AS-C expression. Furthermore, they apply only to the ventral ‘neurogenic’ ectoderm, where the relation between proneural clusters and neuroblast segregation has not yet been worked out.

Situations where only one of a group of equipotent cells undergoes a given change have been first described in the case of insect sensory bristles by Wigglesworth (1940). He proposed that this restriction could be achieved by a simple system of lateral inhibition (rev. in Simpson, 1990). One possible mechanism could be that competent cells compete for a limiting amount of a bristle-inducing substance. The first cell to become SMC will deplete the substance and thereby effectively prevent the surrounding cells from becoming SMCs. An alternative mechanism would be that SMCs actively inhibit their neighbours by producing a specific inhibitory signal. In Drosophila, lateral inhibition has been invoked (Ghysen and Richelie, 1979) to explain some cases of non-autonomy indicating that the decision to form a sensory bristle is not taken by isolated cells, but by groups of cells (Stem, 1954). More recently lateral inhibition has been experimentally demonstrated in the segregation of central neuroblasts in the grassshopper (Doe and Goodman, 1985).

Genetic analyses of the vulval development of the nematode has suggested that lateral inhibition may be more complex than simply an all- or-none effect of the first cell reaching some threshold becoming instantly able to block all its neighbours. In the nematode case, the results indicate that mutual inhibition leads to an unstable situation where any small difference will be self-reinforcing, thus ensuring that only oné cell will ever predominate (Seydoux and Greenwald, 1989). It has been proposed that the singling out of bristle mother cells in the fly might rely on a similar mechanism of balanced mutual inhibition, and that the neurogenic loci would mediate this mechanism (Heitzler and Simpson, 1991).

We have relied on the development of enhancer-trap lines (O’Kane and Gehring, 1987) where the reporter gene lacZ is specifically expressed in the neural precursors and in their progeny (Ghysen and O’Kane, 1989) to assess directly the effect of mutations in six of the seven neurogenic loci on the emergence of these precursors.

Fly strains

The transformant lines A37 and A18 have been described in Ghysen and O’Kane (1989). The line 6727 contains a P[ry+, ftz/lacZF] element inserted just upstream of the transcription initiation site of the elav gene, near the tip of the X chromosome (band IB), and causes lacZ expression in 7 stripes early on, followed by expression in all CNS neurons at later stages. The CyOβE3 balancer is a CyO balancer chromosome which carries an insertion of the ftz/lacZQ element (Hiromi et al. 1985), resulting in a weak striped expression of lacZ in the epidermis and a segmentally repeated expression in a defined subset of neuroblasts in the CNS. The TM3βT8 is a TM3, Sb balancer chromosome with a P element containing the UPHZ50T construct (Hiromi and Gehring, 1987), resulting in a strong expression of lacZ in parasegment 4. All three lines were produced and kindly provided by Y. Hiromi.

The alleles that we used are, for almondex: amx1, for big brain: bib11195, for Delta: Dl9P39, for Enhancer of split: E(spl)RA.7 1, for neuralised: neuKX9, for Notch, either N8, a small deficiency deleting N, or N55eI1, a null mutation of the gene, to generate the germ-line clones (homozygous N8 germ cells fail to form oocytes).

The A18, A37 and B52 insertion chromosomes were multiply marked to facilitate the selection of DI9P39 A18, E(spl)RA.7.1 A37, and neuKX9 B52, recombinant chromosomes. All recombinant chromosomes were balanced with TM3βT8 to allow us to recognize without ambiguity the homozygous mutant embryos.

In the case of big brain, which maps on the second chromosome, we crossed bib/SM females with CyOβE3/+; A37/A37 males. We crossed the F1, bib/CyOβE3; A37/+ males and females, and labelled the F2 embryos.

The mutations amx and N map on the X chromosome, amx is homozygous viable: the neurogenic phenotype is shown only in homozygous embryos derived from homozygous mothers. We crossed amxlamx females with amx/Y’, A37/+ males, and examined their progeny. In the case of Notch (N8 or N55ell), we crossed N/FM4 females with6727/Y; A37/A37 males, selected the N/6727’, A37/+ females, backcrossed them with 6727/Y; A37/A37 males and labelled the F2 embryos.

To obtain embryos lacking maternal as well as zygotic Notch product, we proceeded as follow: N55ell/FM7 females were crossed with ovoD1/Y males. Mitotic recombination was induced in the germ line of the F1N/ovoD1 females by irradiation with X-rays (Müller MG, 100kV, 15 mA, 300 R min-1’, total dose of 1000 R) 48–72 h after egg laying. 500 irradiated N females were selected and crossed with 6727/Y; A37/A37 males and their progeny was labelled. A total of 170 embryos were obtained, 82 of which did not show the 6727 pattern and were consequently N55ell/Y embryos. Of these, 23 were at the appropriate stage (5h –6h of development).

lacZ staining in the embryos

For the X-gal staining, see Ghysen and O’Kane (1989). Immunostainings were done as follows (all steps are at room temperature in microtiter wells). Dechorionated embryos were fixed for 20 min on a rocking platform in a 1:1 mixture of fixative (7 % formaldehyde in PBS) and heptane. The fixative was replaced by cold methanol and agitated strongly for 20 s to remove the vitelline membrane. The embryos were then rinsed 3 times with fresh methanol, and either used immediately or stored at −20°C. The embryos were rehydrated in PBS:methanol (1:1), rinsed three times in PBT (PBS with 0.1% Tween), blocked in PBS+0.5% BSA for 45 min, and incubated overnight in the primary antibody (Promega monoclonal mouse anti-β-galactosidase diluted 1:500 in PBT). After 3 rinses followed by three washes of 20 min each in PBT, the embryos were incubated for 2 h in the presence of the secondary antibody (anti-mouse IgG biotinylated antibody, Vector labs, diluted 1:400 in PBT and preabsorbed for 2h to fixed devitellinized embryos), rinsed three times and washed three times for 20 min each in PBT and incubated in the avidin/biotinylated peroxidase/PBT solution (Vectastain, Vector labs) which had been preincubated for 60 min. After three rinses and three washes of 20 min in PBT, the embryos were transferred in the staining solution (0.05% DAB and 0.03% H2 O2 in PBT). The reaction was followed under the dissection microscope and stopped by several rinses with PBT. The embryos were dehydrated through an ethanol series (70 %, 90 % and 3 times 100%), transferred to xylene and mounted in Canada Balsam:xylene mounting medium. Photographs were taken on Ilford PanF film using a Zeiss Universal microscope equipped with neofluar 40× or 100× objectives.

(1) Early pattern of SMCs in normal and Delta embryos

We limited our analysis to the SMCs that form in the body segments, to the exclusion of the head and terminal region of the abdomen. The pattern of appearance of SMCs in these segments, as detected in the A37 line, is very reproducible (Ghysen and O’Kane, 1989). The first SMC appears early during embryogenesis, at about 5h of development at 25 °C, in the posterior region of each segment (P cell). The emergence of P cells in the different segments is pairrule-wise: Al and A3 first, then A5 and A7. P cells become detectable slightly but consistently later in the other body segments, and even later in A8. At about 5h10 of development, a second SMC forms in the anterior region of each segment (A cell). This second SMC appears in all segments more or less synchronously, at the same dorsoventral level as the P cell (Fig. 1A). Pairs of A and P cells, probably arising from the divison of the early A and P cells (at least in the case of the A pair, unpublished observations) are soon observed (about 5h20 of development), followed by the emergence of one dorsal and one ventral SMC (Fig. 1C). Afterwards, the situation becomes complex as more cells appear at the different positions. After germ band retraction, the cells are arranged in a very regular segmental pattern composed of a dorsal, a lateral and two ventral clusters (Fig. 1E).

Fig. 1.

Emergence of the sensory mother cells (SMCs) pattern in A37 and Df9P39 A37 embryos. (A, B) Emergence of the first SMCs in young embryos (5h10–5h20 of development). (A) In normal embryos, most segments contain one anterior, A, and one posterior, P, cell. The P cell is somewhat obscured by the labelling of a cluster of epidermal cells (arrowheads) around the same position. (B) In DI A37 embryos, clusters of A and P cells are present in all segments. (C, D) Slightly older embryos (5h20–5h30 of development). (C) In normal embryos, pairs of P and A cells are labelled; one dorsal, D, and one ventral, V, SMC are also present. (D) In Dl embryos, large clusters of A and P cells are present; D cells are appearing at the appropriate position but in excessive numbers (arrowheads). (E, F) Late embryos after germ band shortening. (E) Normal embryos show a regular pattern of labelled cells. (F) in Dl embryos, the peripheral nervous system is hypertrophied and disorganized, partly because the ventral region has been entirely converted into neuroblasts (asterisks), which are not labelled in the A37 fine, and partly because of the massive depletion of epidermis. In all panels, anterior is to the left and dorsal to the top. T3=metathoracic segment, A7=seventh abdominal segment.

Fig. 1.

Emergence of the sensory mother cells (SMCs) pattern in A37 and Df9P39 A37 embryos. (A, B) Emergence of the first SMCs in young embryos (5h10–5h20 of development). (A) In normal embryos, most segments contain one anterior, A, and one posterior, P, cell. The P cell is somewhat obscured by the labelling of a cluster of epidermal cells (arrowheads) around the same position. (B) In DI A37 embryos, clusters of A and P cells are present in all segments. (C, D) Slightly older embryos (5h20–5h30 of development). (C) In normal embryos, pairs of P and A cells are labelled; one dorsal, D, and one ventral, V, SMC are also present. (D) In Dl embryos, large clusters of A and P cells are present; D cells are appearing at the appropriate position but in excessive numbers (arrowheads). (E, F) Late embryos after germ band shortening. (E) Normal embryos show a regular pattern of labelled cells. (F) in Dl embryos, the peripheral nervous system is hypertrophied and disorganized, partly because the ventral region has been entirely converted into neuroblasts (asterisks), which are not labelled in the A37 fine, and partly because of the massive depletion of epidermis. In all panels, anterior is to the left and dorsal to the top. T3=metathoracic segment, A7=seventh abdominal segment.

In a DI embryo, SMCs appear first at a position characteristic for the P cells, then at a position that corresponds to that of the normal A cells (Fig. 1B). Thus the earliest SMCs appear in the right sequence (P first, followed by A) at the right place. However, the SMCs do not appear in the right number: several P cells (3–5) are present when the first A cell forms, and several A cells are present before D or V SMCs appear (Fig. 1B). This never occurs in wild-type embryos where clusters of cells at the A and P positions are observed much later, after the appearance of dorsal and ventral mother cells. The pattern becomes more and more abnormal at later times, as new clusters of SMCs presumably corresponding to the D cells form (Fig. 1D), leading after germ band retraction to a disorganized mass of sense cells (Fig. 1F). This sequence is very similar to what we had observed earlier in embryos deficient for N (Ghysen and Dambly-Chaudière, 1990).

(2) All neurogenic mutations have a similar early phenotype

We examined the effect on SMC formation of mutations in six of the seven neurogenic loci: N, DI, E(spl), neu, amx and bib. Whenever possible, we used deficiencies to make sure that we are dealing with a complete loss of zygotic function. In all cases, embryos homozygous for the mutation were identified unam-bigously because the balancer chromosomes contain a fiz-lacZ construct where lacZ is expressed in a specific striped pattern.

One of the mutants, neu, maps relatively close to the lacZ insert in A37, at band 80A on the third chromosome. In this case, we used another enhancertrap line, B52, where lacZ is inserted at band 90B, to facilitate the obtention of recombinants (see Material and Methods). The pattern of expression of lacZ in B52 reveals all SMCs as well as their progeny, much as in the case of A37. Furthermore the sequence of labelling of the early SMCs is identical in A37 and in B52. The patterns of A37 and B52 are not identical, however; in B52 there is a higher level of lacZ expression in the SMCs that correspond to one subset of sense organs, the chordotonal organs (Fig. 2E). This higher level can already be detected in the P cells, which are precursors of chordotonal organs, relative to the A cells, which are precursors of external sense organs (Ghysen and O’Kane, 1989)

Fig. 2.

Early SMC pattern in the different neurogenic mutants. (A) Notch; (B) almondex; (C) Enhancer of split; (D) bigbrain; (E) neuralized. In all cases, several A and P cells are present instead of the normal single ones (compare with the wild-type embryo at the same stage in Fig. 1A). The embryos were stained with X-Gal, which explains the lower resolution of the labelling.

Fig. 2.

Early SMC pattern in the different neurogenic mutants. (A) Notch; (B) almondex; (C) Enhancer of split; (D) bigbrain; (E) neuralized. In all cases, several A and P cells are present instead of the normal single ones (compare with the wild-type embryo at the same stage in Fig. 1A). The embryos were stained with X-Gal, which explains the lower resolution of the labelling.

Fig. 2 shows that all mutants display the characteristic phenotype illustrated Fig. 1B in the case of DI, where the D and V precursors have not yet appeared, yet the A and P pairs have become clusters of 5–8 cells. The occurrence of this phenotype in all cases allows us to extend to all neurogenic loci the conclusions that SMCs form in the right sequence and at the right place, but in abnormal numbers. In mutants with weaker neurogenic phenotypes, such as amx and bib (Fig. 2B, D), the early abnormality of the pattern is milder than in the more extreme mutants. In all cases, however, the anterior and posterior clusters count more than the normal two cells before the dorsal and ventral SMCs can be detected.

(3) The proper specification of SMCs does not depend on the neurogenic loci

Differences between types of sense organs are already manifest at the SMC stage. For example, the genes cut and poxn are specifically expressed in the SMCs that will produce external sense organs and polyinnervated sense organs, respectively (Blochlinger et al. 1990; Bopp et al. 1989; Dambly-Chaudière, unpublished). Some enhancer-trap lines are also specific for subtypes of sense organs. In the line A18, for example, lacZ is expressed in only a few sense organs in each segment: the five lateral chordotonal organs in the abdominal segments, and their thoracic homologs (Ghysen and O’Kane, 1989). We have used this line to examine if the large numbers of SMCs that are formed in neurogenic mutants are unspecified or inappropriately specified sense organ precursors. In a normal A18 embryo, lacZ expression is initially confined to the P pair. In neurogenic mutants, the specificity of A18 expression is maintained, in that the cells of the posterior cluster all express lacZ, while the anterior cells remain silent (Fig. 3A, B). We conclude that the neurogenic mutants do not lead to the formation of large numbers of undefined SMCs, but rather that they result in the overproduction of defined SMCs.

Fig. 3.

The A and P clusters retain their specificity in Dl embryos, as measured by the expression of lacZ in the line A18 which is specific of the P cells and their progeny in normal embryos. (A) Only the P cluster is labelled in young Dl9P39 A18 embryos (same stage as in Fig. 1B). (B) The restriction of A18 expression is maintained in late Dl9P39 A18 embryos (compare with Fig. 1F).

Fig. 3.

The A and P clusters retain their specificity in Dl embryos, as measured by the expression of lacZ in the line A18 which is specific of the P cells and their progeny in normal embryos. (A) Only the P cluster is labelled in young Dl9P39 A18 embryos (same stage as in Fig. 1B). (B) The restriction of A18 expression is maintained in late Dl9P39 A18 embryos (compare with Fig. 1F).

(4) The proper localization of the SMCs does not depend on the neurogenic loci

It has been shown that the phenotype of all neurogenic mutants except bib is stronger when both the maternal and zygotic activities are eliminated than when only the zygotic expression is absent (Jiménez and Campos-Ortega, 1982; Vâssin et al. 1985; Knust et al. 1987). The existence of a maternal contribution implies that the phenotype assessed in homozygous mutant embryos does not reflect the complete loss of function of the gene. It is conceivable that, in the absence of the maternal contribution, SMCs would appear in a disorganized pattern and would not form recognizable A and P clusters. This possibility is ruled out in the case of the amx embryos, which derive from homozygous amx mothers. However, amx is probably a leaky mutation, and therefore the result is not conclusive. We therefore examined the phenotype of a null N mutation after removing the maternal contribution.

Embryos lacking all maternal N product were obtained by inducing, by mitotic recombination, homozygous N/N clones in the germ line of N/ovoD1 females (This experiment was performed with the help of J. F. de Celis in the laboratory of A. Garcia-Bellido). Unirradiated N /ovoD1 females do not lay eggs because ovoD1 dominantly prevents oocyte formation. All eggs laid by irradiated females necessarily originate from germ cells that have become ovo+, and therefore N/N, due to somatic recombination. Irradiated females were crossed with A37 males carrying a ftz-lac construct on their X chromosome, in order to distinguish N/+ female progeny from the experimental N males (for details, see Material and Methods)

The N embryos displayed an early pattern that was indistinguishable from that of N embryos where the maternal contribution had not been removed (Fig. 4B, compare with Fig. 2A). Thus, in the complete absence of N product, the SMCs still appear at the appropriate locations and form well-defined P and A clusters, demonstrating that the activity of N is not required to set up the early pattern of proneural competence. Given the similar behavior of the different neurogenic mutants, we assume that this conclusion can be extended to the whole class of neurogenic loci.

Fig. 4.

Notch phenotype after removal of the maternal contribution. (A, C, E) The first A and P SMCs appear in an A37 embryo (A) at the time when the stomodaeum is well formed (arrowheads in E), the posterior midgut is acquiring a complicated shape (arrows in E) and the pole cells have migrated laterally. (B, D, F) In a Notch embryo derived from homozygous Notch germ cells, clusters of A and P cells are present peripherally at a developmental stage when the stomodaeum is just invaginating (arrowheads in F) and the pole cells are still in the sagittal plane (arrows in F). (A, B) Lateral focal plane; (C, D) sagittal focal plane; (E, F) outline of the stomodaeum and gut in the embryos shown in C and D.

Fig. 4.

Notch phenotype after removal of the maternal contribution. (A, C, E) The first A and P SMCs appear in an A37 embryo (A) at the time when the stomodaeum is well formed (arrowheads in E), the posterior midgut is acquiring a complicated shape (arrows in E) and the pole cells have migrated laterally. (B, D, F) In a Notch embryo derived from homozygous Notch germ cells, clusters of A and P cells are present peripherally at a developmental stage when the stomodaeum is just invaginating (arrowheads in F) and the pole cells are still in the sagittal plane (arrows in F). (A, B) Lateral focal plane; (C, D) sagittal focal plane; (E, F) outline of the stomodaeum and gut in the embryos shown in C and D.

(5) SMCs appear in an ordered manner within the cluster, in neurogenic mutants

The previous results leave little doubt that the neurogenic loci are required for the restriction to a single cell, the SMC, of the proneural capabilities shared by all the cells of the proneural cluster. They throw no light, however, on the question of why that particular cell was chosen. In the imaginai discs, the SMC forms at a reproducible position within each cluster, usually near the center of the cluster, except in clusters that produce more than one SMC (Cubas et al. 1991; Skeath and Carroll, 1991). In the embryo, the SMC also originates at or near the center of the cluster of AS-C expression (M. Ruiz-Gomez, personal communication).

Are the neurogenic loci involved in defining the position where a SMC should form within the proneural cluster? We have adressed this question by following in detail the emergence of the posterior cluster of SMCs in mutant embryos. If the neurogenic loci had a role in patterning the clusters, mutants might show one of two possible results. One possibility is that all cells of the cluster become simultaneously SMCs. The other possibility is that minor differences in the rate of expression of AS-C or other factors will cause the different cells of the cluster to become SMCs at different times, in a random sequence.

Our results (Fig. 5) show that neither of the two possibilities occurs. At early times, we often find small clusters of 2–3 labelled cells, suggesting that all competent cells do not become simultaneously SMCs (Fig. 5A). These clusters increase progressively in size up to a final size of 8–12 cells. The clusters are always composed of contiguous cells (Fig. 5A–F), indicating that the conversion of competent cells into SMCs does not occur at random. In late clusters, all cells are strongly labelled (Fig. 5E). In younger (smaller) clusters, which often contain weakly labelled as well as strongly labelled SMCs, we observe that the weakly labelled cells (arrowheads in Figs 5A,C) are never at the center of the cluster. Taken together, these observations suggest that the SMCs appear in an ordered manner, more peripheral SMCs forming later than more central ones. The same pattern was observed in the other mutants examined in detail (data not shown in the case of bib, E(spl) and amx) as well as in the embryos that lack both maternal and zygotic N product (Fig. 5B).

Fig. 5.

Morphology of the SMC clusters in the abdominal regions of neurogenic mutants. (A, C, E) In Dl A37 embryos, the first SMCs appear as small clusters of 2–3 neighbouring cells (A), increase in size to 5–6 cells (C) and up to 8–12 cells at the moment the ventral V cells appear (E). The SMCs are always adjacent within a cluster. The most peripheral cells are often weakly labelled (arrowheads) when the cluster is expanding (A, C) but not when the clusters have reached their final size (E), supporting the conclusion that the increase in size of the clusters is due to the emergence of additional SMCs around the first SMCs to be formed (see text). (B) In a N; A37 embryo devoid of maternal Notch product, the morphology of the clusters is similar to what is found in the Delta embryo. (D, F) In a B52 neu embryo, the development of the clusters is also progressive, more central cells appearing before the more peripheral ones. (D) In the three P clusters shown, the central nucleus (arrowed) is more darkly labelled than in the surrounding cells, presumably because the central cell became SMC before the others. The large structure arrowed in (F) probably corresponds to a SMC entering mitosis.

Fig. 5.

Morphology of the SMC clusters in the abdominal regions of neurogenic mutants. (A, C, E) In Dl A37 embryos, the first SMCs appear as small clusters of 2–3 neighbouring cells (A), increase in size to 5–6 cells (C) and up to 8–12 cells at the moment the ventral V cells appear (E). The SMCs are always adjacent within a cluster. The most peripheral cells are often weakly labelled (arrowheads) when the cluster is expanding (A, C) but not when the clusters have reached their final size (E), supporting the conclusion that the increase in size of the clusters is due to the emergence of additional SMCs around the first SMCs to be formed (see text). (B) In a N; A37 embryo devoid of maternal Notch product, the morphology of the clusters is similar to what is found in the Delta embryo. (D, F) In a B52 neu embryo, the development of the clusters is also progressive, more central cells appearing before the more peripheral ones. (D) In the three P clusters shown, the central nucleus (arrowed) is more darkly labelled than in the surrounding cells, presumably because the central cell became SMC before the others. The large structure arrowed in (F) probably corresponds to a SMC entering mitosis.

(6) SMCs appear earlier in the absence of N function

If the neurogenic loci were required only after the SMC has formed, to establish a region of inhibition around that cell, there is no reason why the time at which the first SMCs form should be affected in neurogenic mutants. However, if these loci were involved in the process by which one cell is singled out from among the others, then the time of SMC formation might be modified in mutants. We compared the time of appearance of the P cells in normal embryos and in N mutant embryos derived from N maternal germ cells (Fig.4).

We assessed the age of embryos by relying on the development of their proctodaeum and stomodaeum, two ectodermal derivatives which make their appearance before the onset of neurogenesis. The development of both structures is essentially normal except in later stages (more than 8h of development) when spatial relations are upset by other abnormalities (Poulson, 1940). In normal embryos, the first A and P cells appear at a time when the proctodaeum and stomodaeum are already well formed (Fig. 4C,E). Among the N embryos that were obtained from N germ cells, only two were at the early stage where clusters of A and P cells are just appearing (Fig. 4B). In both embryos the proctodaeum was still a pouch containing the pole cells, and the stomodaeum was just beginning to invaginate (Fig. 4D, F). SMCs are never observed at this early stage in normal embryos. We conclude that the very first SMCs appear significantly earlier in the complete absence of N product.

The neurogenic loci are not required for the localization and specification of the proneural clusters

Genetic, cellular and molecular studies strongly indicate that the accurate location of a sense organ results from a series of steps that progressively determine the position. The process is initiated by the proneural genes (achaete-scute, daughterless and presumably others) which define groups of competent cells in specific regions of the ectoderm. This ‘proneural pattern’ is then refined in a ‘sensory mother cell pattern’, where one cell from each cluster has been selected to become the sensory mother cell. Our results support the idea that the neurogenic mutants alter this second step.

Previous experiments had shown that the additional SMCs in neurogenic mutants form in the same regions where proneural clusters normally arise in wild type embryos (Cabrera, 1990; Brand and Campos-Ortega, 1989). However, in these experiments only N and DI, the two loci coding for transmembrane proteins, were examined. Furthermore, the possibility that the phenotype observed corresponds only to a partial loss of function remained unexplored, since the influence of the maternal contribution on the pattern of SMCs was either not assessed (Cabrera, 1990) or assessed at the level of the final differentiated phenotype (Brand and Campos-Ortega, 1989), from which it is not possible to infer whether or not the early SMCs had formed at the appropriate positions.

We observed that all the six neurogenic mutations examined in this paper (amx, bib, DI, E(spl), neu and N) cause the same early phenotype: several SMCs appear at the positions where only a pair should form. This phenotype is also observed in N embryos where the maternal contribution had been removed. Thus, even in the complete absence of N product the formation of the early SMCs occurs in the appropriate sequence and at the appropriate place, only the numbers are abnormal. The results with A18 and B52 also show that the neurogenic loci are not involved in the system that confers different specificities to SMCs formed at different positions, in agreement with the observations of Brand and Campos-Ortega (1989).

We conclude that the neurogenic loci are required to limit the number of competent cells that will become SMCs, but play no role in the early steps that localize where proneural genes are expressed and define which type of SMC will form.

Lateral inhibition versus mutual inhibition

Restricting the number of competent cells that will become SMCs might be achieved in two ways. One way is that once a SMC has formed, it produces an inhibitory signal mediated by the neurogenic loci. Accordingly, the products of the loci would play no role before the first SMC is formed; only then would they begin to function and prevent the neighbouring cells from adopting the same fate. An alternative possibility is that the role of the neurogenic loci is to set up a system of mutual inhibition which is already active before the SMC emerges. Provided that there is a feedback mechanism whereby the level of inhibiting signal emitted by one cell is inversely related to the level of inhibitory signal received by that cell, then the system will amplify small differences such that one cell will become fully inhibitory while its neighbours will become fully inhibited (Heitzler and Simpson, 1991). In this way, the commitment of one cell and the inhibition of its neighbours would be two aspects of the same process.

If the neurogenic loci were involved in setting up the inhibition after the SMC has formed, as in the first possibility, then the lack of function of these genes should have no effect whatsoever on the time of emergence of the first SMCs. If, on the contrary, the appearance of a SMC is preceded by a period of mutual inhibition, then removing this inhibition would result in an earlier appearance of the first SMCs. We observed that, in the complete absence of N product, the first SMCs form earlier than in normal embryos. This finding is entirely consistent with the proposal that the neurogenic loci are involved in setting up a system of competitive inhibition such that small differences in the level of competence are gradually transformed into an all- or-none difference between one maximally competent cell, the SMC, and a number of minimally competent cells which will be reduced to an epidermal fate.

Structure of the equivalence groups and choice of the SMC

The conversion of a proneural cell into a SMC is related to the accumulation of high levels of proneural gene products, both in the adult and in the larval PNS (Cabrera, 1990; Cubas et al. 1991; Skeath and Carroll, 1991; M. Ruiz-Gomez, personal communication). However, the mechanism underlying the choice of one particular cell among the equivalent cells of the proneural cluster is still poorly understood. This mechanism has been best studied in the imaginai wing disc by examining chimaeric proneural clusters comprising cells with variable doses of the genes thought to be involved in the choice. Cubas et al. (1991) have shown that the dose of AS-C is important, since in clusters comprising cells with one or two doses of AS-C, the SMC preferentially arises from one of the cells with two doses. Likewise de Celis et al. (1991) have observed that proneural cells with fewer doses of N than their neighbours become preferentially SMCs. Heitzler and Simpson (1991) conclude from their extensive analysis that the choice of a SMC depends on the dose of N and of Dl in an antagonistic manner: cells with less N or more Dl than their neighbours have a higher probability of becoming SMCs. They propose that the N product is the receptor of the inhibitory signal, while Dl would be the inhibiting ligand. Taken together, these results indicate that the choice of which cell will become the SMC depends on the balance between the accumulation of proneural gene products and the inhibitory influence of the surrounding cells. Gene dosage effects that increase the amount of AS-C products in a cell, or decrease its sensitivity to inhibition, or increase its inhibitory effect, will all increase the chance that this cell will become SMC.

We have observed that, in neurogenic mutants, the SMCs appear in an orderly manner so that the cluster expands progressively from 2–3 to 8–12 adjacent cells. Thus even in the strongest neurogenic mutants, the cells in the cluster may be equivalent in the sense that they are all capable of becoming a SMC, but they are nevertheless not identical since they become SMCs at different times.

This indicates that independently of the neurogenic loci, there exists a mechanism that ensures that some of the cells of the proneural cluster, probably those at the center, are the most likely to become SMCs. An obvious possibility is that this difference between competent cells is due to differences in the level of expression of the proneural genes in the different cells of the cluster. Indeed if the AS-C genes are controlled by combinations of regionally expressed genes, it is very likely that the rate of activation will be highest near the center of the cluster. The predisposition of the central cells to become the first SMCs can be offset if the dosage of the AS-C or neurogenic genes is altered in some of the cells, leading to the observed bias in the cell’s choice of fate in mosaics of cells bearing different doses of N and DI genes (Heitzler and Simpson, 1991) or of AS-C genes (de Celis et al. 1991). On the other hand if there is no such dosage alteration, then the central cell is the most likely to become SMC, as is indeed the case in normal embryos and wing discs (M. Ruiz-Gomez, personal communication, Cubas et al. 1991).

We are grateful to the laboratory of A. Garcia-Bellido, and particularly to J. F. de Celis, for hospitality and help with the generation of the germ-line clones. We thank Y. Hiromi for giving the fac-expressing balancer strains that were essential for this work, P. Simpson, M. Ruiz-Gomez and J. Modolell for communicating unpublished results, and J. Modolell for comments on the manuscript. This work was supported by a contract between the Belgian government and the University of Brussels (Actions de Recherche Concertée) and by a grant from the HFSPO (Human Frontier Science Programme Organization). A. Goriely is a fellow of the IRSIA (Institut pour 1’Encouragement de la Recherche Scientifique dans l’industrie et l’Agriculture) and was supported by an EMBO fellowship during her stay in Madrid. A. Ghysen is chercheur qualifié of the FNRS (Fonds National de la Recherche Scientifique, Belgium).

Blochlinger
,
K. B.
,
Bodmer
,
R.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1990
).
Patterns of expression of Cut, a protein required for external sensory organ development in wildtype and cut mutant Drosophila embryos
.
Genes Dev
.
4
,
1322
1331
.
Bopp
,
D.
,
Jamet
,
E.
,
Baumgartner
,
S.
,
Burri
,
M.
and
Noll
,
M.
(
1989
).
Isolation of two tissue-specific Drosophila paired box genes, Pox meso and Pox neuro
.
EMBO J
.
8
,
3447
3457
.
Brand
,
M.
and
Campos-Ortega
,
J. A.
(
1989
).
Two groups of interrelated genes regulate early neurogenesis in Drosophila melanogaster
.
Roux’s Arch. devl. Biol
.
197
,
457
470
.
Cabrera
,
C. V.
(
1990
).
Lateral inhibition and cell fate during neurogenesis in Drosophila: the interactions between scute, Notch and Delta
.
Development
109
,
733
742
.
Cabrera
,
C. V.
,
Martinez-Arias
,
A.
and
Bate
,
M.
(
1987
).
The expression of three members of the achaete-scute gene complex correlates with neuroblast segregation in Drosophila
.
Cell
50
,
425
433
.
Campos-Ortega
,
J. A.
(
1988
).
Cellular interactions during early neurogenesis of Drosophila melanogaster
.
Trends Neurosci
.
11
,
400
405
.
Cubas
,
P.
,
De Celis
,
J. F.
,
Campuzano
,
S.
and
Modolell
,
J.
(
1991
).
Proneural clusters of achaete-scute expression and the generation of sensory organs in the Drosophila imaginai wing disc
.
Genes Dev
.
5
,
996
1008
.
De Celis
,
J. F.
,
Mari-Beffa
,
M.
and
Garcia-Bellido
,
A.
(
1991
).
Cell autonomous role of the Notch gene, an epidermal growth factor homolog, in sensory organ differentiation in Drosophila
.
Proc. natn. Acad. Sci. U.S.A
.
88
,
632
636
.
Doe
,
C. Q.
and
Goodman
,
C. S.
(
1985
).
Early events in insect neurogenesis. II: the role of cell interactions and cell lineage in the determination of neuronal precursor cells
.
Devl Biol
.
111
,
206
219
.
Ghysen
,
A.
and
Dambly-ChaudièRe
,
C.
(
1989
).
The genesis of Drosophila Peripheral Nervous System
.
Trends Genet
.
5
,
251
255
.
Ghysen
,
A.
and
Dambly-ChaudièRe
,
C.
(
1990
).
Early events in the development of Drosophila peripheral nervous system
.
J. Physiol. (Paris)
84
,
11
20
.
Ghysen
,
A.
and
O’Kane
,
C.
(
1989
).
Detection of neural enhancer-like elements in the genome of Drosophila
.
Development
105
,
35
52
.
Ghysen
,
A.
and
Richelle
,
J.
(
1979
).
Bristle determination and pattern formation in Drosophila. II. The ac-sc locus
.
Devl Biol
.
70
,
438
452
.
Hartenstein
,
V.
and
Campos-Ortega
,
J. A.
(
1986
).
The peripheral nervous system of mutants of early neurogenesis in Drosophila melanogaster
.
Roux’s Arch. devl. Biol
.
195
,
210
221
.
Hartley
,
D. A.
,
Preiss
,
A.
and
Artavanis-Tsakonas
,
S.
(
1988
).
A deduced gene product from the Drosophila neurogenic locus, Enhancer-of-split, shows homology to mammalian G-protein B subunit
.
Cell
55
,
785
795
.
Heitzler
,
P.
and
Simpson
,
P.
(
1991
).
The choice of cell fate in the epidermis of Drosophila
.
Cell
64
,
1083
1092
.
Hiromi
,
Y.
and
Gehring
,
W. J.
(
1987
).
Regulation and function of the Drosophila segmentation gene fushi tarazu
.
Cell
50
,
963
974
.
Hiromi
,
Y.
,
Kuroiwa
,
A.
and
Gehring
,
W. J.
(
1985
).
Control elements of the Drosophila segmentation gene fushi tarazu
.
Cell
43
,
603
613
.
Huang
,
F.
,
Dambly-ChaudièRe
,
C.
and
Ghysen
,
A.
(
1991
).
The emergence of sense organs in the wing disc of Drosophila
.
Development
111
,
1087
1095
.
Jiménez
,
C.
and
Campos-Ortega
,
J. A.
(
1982
).
Maternal effects of zygotic mutants affecting early neurogenesis in Drosophila. Wilhelm Roux’s Arch. devl
.
Biol
.
191
,
191
201
.
Kidd
,
S.
,
Kelley
,
M. R.
and
Young
,
M. W.
(
1986
).
Sequence of the Notch locus of Drosophila melanogaster: relationship of the encoded protein to mammalian clotting and growth factors
.
Molec. cell. Biol
.
6
,
3094
3108
.
Klämbt
,
C.
,
Knust
,
E.
,
Tietze
,
K.
and
Campos-Ortega
,
J. A.
(
1989
).
Closely related transcripts encoded by the neurogenic gene complex Enhancer of split of Drosophila melanogaster
.
EMBO J
.
8
,
203
210
.
Knüst
,
E.
,
Bremer
,
K. AL.
,
Vassin
,
H.
,
Ziemer
,
A.
,
Tepass
,
U.
and
Campos-Ortega
,
J. A.
(
1987a
).
The Enhancer of Split locus and neurogenesis in Drosophila melanogaster
.
Devl Biol
.
122
,
262
273
.
Kopczynsm
,
C. C.
,
Alton
,
A. K.
,
Fechtel
,
K.
,
Kooh
,
P. J.
and
Muskavitch
,
M. A. T.
(
1988
).
Delta, a Drosophila neurogenic gene, is transcriptionally complex and encodes a protein related to blood coagulation factors and epidermal growth factor of vertebrates
.
Genes Dev
.
2
,
1723
1735
.
Lehmann
,
R.
,
JiméNez
,
F.
,
Dietrich
,
U.
and
Campos-Ortega
,
J. A.
(
1983
).
On the phenotype and development of mutants of early neurogenesis in Drosophila melanogaster. Wilhelm Roux’s Arch
.
devl. Biol
.
192
,
62
74
.
O’Kane
,
C. J.
and
Gehring
,
W.
(
1987
).
Detection in situ of genomic regulatory elements in Drosophila
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
9123
9127
.
Poulson
,
D. F.
(
1940
).
The effect of certain X-chromosome deficiencies on the embryonic development of Drosophila melanogaster
.
J. exp. Zool
.
83
,
271
318
.
Seydoux
,
G.
and
Greenwald
,
I.
(
1989
).
Cell autonomy of lin-12 function in a cell fate decision in C. elegans
.
Cell
57
,
1237
1245
.
Simpson
,
P.
(
1990
).
Lateral inhibition and development of sensory bristles of the adult peripheral nervous system of Drosophila
.
Development
109
,
509
519
.
Skeath
,
J. B.
and
Carroll
,
S. B.
(
1991
).
Regulation of achaete-scute gene expression and sensory organ pattern formation in the Drosophila wing
.
Genes Dev
.
5
,
984
995
.
Stern
,
C.
(
1954
).
Two or three bristles
.
Am. Scientist
42
,
213
247
.
Vässin
,
H.
,
Bremer
,
K. A.
,
Knust
,
E.
and
Campos-Ortega
,
J. A.
(
1987
).
The neurogenic gene Delta of Drosophila melanogaster is expressed in neurogenic territories and encodes a putative transmembrane protein with EGF-like repeats
.
EMBO J
.
6
,
3431
3440
.
Wharton
,
K. A.
,
Johansen
,
K. M.
,
Xu
,
T.
and
Artavanis-Tsakonas
,
S.
(
1985
).
Nucleotide sequence from the neurogenic locus Notch implies a gene product that shares homology with proteins containing EGF-like repeats
.
Cell
.
43
,
567
581
.
Wigglesworth
,
V. B.
(
1940
).
Local and general factors in the development of ‘pattern’ in Rhodnius prolixus (Hemiptera)
.
J. exp. Biol
.
17
,
180
200
.