Search

In insects, neurons in the central nervous system (CNS) are generated by progenitor cells called neuroblasts derived from a region of the ectoderm called the neurogenic region or neuroectoderm. In the neuroectoderm of Drosophila melanogaster, neighboring cells take on one of two alternative fates and develop either as neuroblasts or as epidermoblasts (progenitor cells of the epidermis). The neuroblasts move to deeper levels of the embryo to build up the central neural primordium, whereas the epidermoblasts remain at the surface to build up part of the epidermal sheath. The peri­ pheral nervous system (PNS) of insects develops from progenitor cells located within the epidermis. Thus, development of the PNS involves another choice by the epidermal cells between neural and nonneural fates, i.e. to develop as sensory progenitor cells versus non sensory epidermal cells. As is discussed in this review, many of the molecular mechanisms involved in the cell fate choices are shared in the development of the CNS and PNS. The analysis of the cellular decisions that lead to the formation of the

Die Entwicklungsbiologie beschaftigt sich mit zwei Hauptfragen: der Entstehung von Zelldiversitat, d.h., wie sich die verschiedenen, in einem multizellularen Organismus vorkommenden Zelltypen entwickeln, und der Musterbildung, d.h., wie Zellen sich gruppieren, um Gewebe und Organe zu bilden. Zwar konnen diese Fragen auf Grund ihrer Komplexitat noch nicht umfassend beantwortet werden, Entwicklungsbiologen haben aber einige der beteiligten Grundelemente und Mechanismen weitgehend aufgeklart.

The Drosophila achaete-scute complex consists of four genes encoding transcription factors of the bHLH family. Due to their intricate organization, these genes have occupied geneticists and developmental biologists for many years. Here, genetic studies on the complex are discussed from a historical point of view.

Cells of the neurogenic ectoderm of insects have to decide between a neural and an epidermal fate. In Drosophila, this decision id mediated by cellular interactions. The products of two different groups of genes, i.e., the neurogenic genes and the genes of the achaete-scute complex and daughterless, seem to provide the molecular basis for the elements of a signal chain that permits the commitment of the cells to a given fate.

The development of the rhabdomeric pattern in the compound eye ofDrosophila has been studied using combined transplantation and electron microscope techniques. In a first series of experiments eye imaginal discs of increasing age were implanted into larvae ready to pupate, thus losing variable amounts of the normal time for development. A sequence of differentiative abilities was found in the metamorphosed test pieces. As far as the photoreceptor cells are concerned, the most prominent steps of this sequence are: ability to form groups with other similar elements, anatomical polarization of microvilli, establishment of the rhabdomeric pattern and formation of an equator line. The stability of determination of the equator line was tested in a second experimental series. Fragment of different topographical origin within the mature eye anlage were brought to metamorphosis by implantation into larvae ready to pupate. It was found that an equator line differentiates only in those pieces which according to the published anlage maps contain the prospective equator region prior to metamorphosis. The mitotic abilities of implanted eye imaginal discs were investigated by means of “in vitro”3H-thymidine pulse-labelling and light microscope autoradiography of the differentiated test pieces. During the third larval stage the eye anlage is traversed by two consecutive mitotic waves, each one of them producing different categories of receptor cells. The first, anterior wave predominantly produces cells oriented toward the poles of the eye within the ommatidia, while the second, posterior wave gives rise to elements exclusively in an equatorial position. The dynamics of this proliferation are discussed in relation to the findings in the implantation experiments. Silver-grain counts support the possibility that at least two successive cell divisions occur in the eye anlage between labeling with tritiated thymidine and beginning of morphological differentiation. The relevance of this finding for the understanding of the concept of acquisition of competence is discussed.

Genetic eye mosaics ofDrosophila melanogaster have been studied by means of anatomical techniques. Using different cell markers it was found that the ommatidia at the boundaries between phenotypes are composed of cells belonging to different clones. Therefore, the formation of an individual ommatidium does not obey a mechanism based on a common clonal origin of its constituent elements. A statistical analysis of mosaic ommatidia shows that there is a significant tendency for the receptor cellsR2-R5 on the one hand and the receptor cellsR1, R6 andR7 on the other to belong to the same cell clone. The implications of these findings are discussed.

The size of the neurogenic region ofDrosophila melanogaster is under the control of several genes of zygotic expression. Lack of function from any of those genes produces an increase of the size of the neurogenic region at the expense of the epidermal anlage. However, differences exist in the extent of neuralisation achieved by each of the genetic loci upon mutation. The present results show that in the case ofN andmam phenotype differences are due to different contributions of maternal gene expression. This could be shown by studying the phenotype which appeared in mutant embryos when the oocytes developed from homozygous mutant precursor cells. Clones of mutant cells were induced in the germ line of females heterozygous for the neurogenic mutationin trans over germ line dependent, dominant female sterile mutations. After removing maternal information the phenotype ofN andmam mutants became identical in both cases. Furthermore maternal information fromN+ was found to be necessary for viability of the wildtype.

We describe a set of cells in the central nervous system of theDrosophila embryo which are restricted to the thoracic ganglia in the wildtype. Taking these cells as indication of thoracic identity, we find that the ventral cord of embryos homozygous mutant for different bithorax functions and for Polycomb undergoes homoeotic transformations equivalent to those observed in the larval cuticle.

Optic cartridges of the first synaptic region of the fly's visual system are interconnected by a set of collaterals derived from the L4 monopolar cells. These neurones are the only efferent monopolar cells to contact elements in more than one cartridge. Electron microscopy of Golgi-impregnated cells has shown that sets of distal (outer) processes of L4 contact, and are postsynaptic to, the α profiles of lamina intrinsic cells at single cartridges. L4's proximal (inner) collaterals, situated at the inner face of the lamina, contact the receptor terminals and the radial monopolar cells (L1 and L2) of its own and of two adjacent cartridges. One L4 meets components of 6 other L4 cells of 6 surrounding cartridges. Conventionally fixed material shows that the proximal collaterals of L4 are presynaptic to radial monopolar cells and some receptor endings as well as being pre-and postsynaptic to the other L4 neuronal components to which they extend, or which extend to them. The numbers of the distal processes postsynaptic to α fibres vary across the eye. There is a graded increase from back to front and from the dorsal and ventral poles of the eye towards its equator.

Only three of the five types of monopolar cells which are present in each cartridge of the lamina have synaptic connections with receptor endings (R1–R6). Due to the distribution of their dendrites two of these (L1 and L2) contact the whole length of the six receptor endings of their cartridge whereas the third type (L3) contacts only their outer 2/3rds. Although these three cells may function as the second order neurons of the neurosuperposition eye, their anatomical relationship imply functional differences between them.

The generalogical relationships of photoreceptor cells within the compound eye ofDrosophila have been studied using cell labelling, with either3H-thymidine or recessive mutations, during the third larval stage. It has been found that photoreceptor and secondary pigment cells arise from different precursor cells. Under the present experimental conditions, precursors of receptor cells give rise to about 8 elements which differentiate as R cells of two different groups. One of the cells differentiates as R7 and the remaining as any one of the R1 to R6. The last cells behave initially as equivalent, and can differentiate within the same or within different, but neighbouring, ommatidia. The class of R1 to R6 cell in which each one of these elements differentiates, seems to depend on the time of its origin. The implications of these findings for the formation of the ommatidial pattern are discussed.

sevLY3,the only existing allele at thesev locus (1–33,2±0,2), behaves as strongly hypomorph or even as amorph. Ommatidia in asev compound eye have only seven receptor cells, the position of the R7 pattern element being vacant. Various criteria showing that the missing cell is R7 have been verified. These include (i) anatomical characteristics ofsev ommatidia; (ii) behaviour of central R cells insev rdgB double mutants; (iii) medullary projection of central R cell axons; and (iv) mitotic pattern ofsev imaginal discs. The analysis of morphogeneticsev-sev+ mosaics has shown thatsev is expressed autonomously by R7 cells, indicating that thesev phenotype is not due to asev genotype of ommatidial pattern elements other than R7. The study of third instarsev imaginal discs has not brought any direct evidence for death of clustered presumptive R7 cells; however, clonal analysis of the developingsev compound eye has given evidence of developmental parameters comparable to those ofsev+, therefore favouring the hypothesis that R7 cells die insev mutants. On the other hand,sev+ seems to be required for the determination of the R7 cells, since thesev phenotype cannot be uncovered during the last mitoses of heterozygous mutant cells.

Mutations in genes involved in essential aspects of central nervous system development in Drosophila melanogaster are expected to be lethal. Thus, when searching for neurogenic mutants attention should be focused on embryonic lethal point mutants1, for many of these might affect neural development. However, this approach can be very time consuming, for the location of neurogenic genes is unknown. A more convenient approach, which allows a faster screening of the genome, is to use relatively small chromosome deletions2,3 to determine whether the lack of a definite part of the genome affects neurogenesis. Once any region producing an interesting neural phenotype is found, it can be further analysed by the use of smaller deletions or point lethal mutants mapping within it, until the gene(s) responsible can be more precisely localised1,4. We report here on a region of the Drosophila genome which has been found necessary for normal neurogenesis.

The neurogenic genes of Drosophila melanogaster are involved in the decision of ectodermal cells to take on a neural or an epidermal fate. We present evidence in support of the notion that six of the neurogenic genes are functionally related. We studied the phenotype of embryos lacking one of the neurogenic genes in the presence of an increased dosage of the wild-type allele of another neurogenic gene. Our analysis also included the Hairless locus, whose function is related to that of the neurogenic genes, as well as to many other genes. The effects observed were asymmetric in that triploidy for a given gene modified the phenotype of loss of the function of another gene, but triploidy of the latter gene did not modify the phenotype of loss of the function of the former gene. These asymmetries allowed us to establish a polarity of gene interactions, as well as to order the genes according to the assumed ability of some of them to modify the activity of others. In this sequence, almondex is the first link and Enhancer of split the last one. Our evidence suggests that the function of big brain is independent of the function of the other six. The consequences of this arrangement for the commitment of ectodermal cells are discussed.

Neurogenic mutations have been found to cause the neuralization of certain regions of the ectoderm and yet to permit normal development of the remaining embryonic cells. Thus, it seems that the activity of the wild-type alleles of these genes is dispensable in a considerable fraction of the embryo during wild-type development. This effect might be a consequence of the cells' position within the embryo; alternatively, it might be independent of the position but be due rather to the genetic activity experienced by the cells previous to their commitment. The results described in this paper indicate that genes controlling patterning along the embryonic dorso-ventral perimeter (dorsal and Toll) are epistatic to genes controlling neurogenesis, their activity deciding which ectodermal cells are susceptible to neurogenesis. Using alleles with low expressivity, evidence was obtained showing that the tracheal placodes define the boundary of the territory which has neurogenic abilities at thoracic and abdominal levels.

Five regions of the compound eye have been found to be preferential boundaries for clones of labelledMinute + cells, and to act restrictively on the growth of cell clones after a given developmental stage. One of these regions is topographically related to the line of pattern inversion existing at the level of the equator. The results of experiments showing independency of origin of restriction lines and line of pattern inversion are reported.

CELLULAR DIVERSITY in an organ might arise from differential mitoses within a prefixed, genetically determined cell lineage or through the reading of positional values by equipotent cells in a morphogenetic field1. We have studied this problem in the developing eye imaginal disc of Drosophilia, in particular in the cells giving rise to the eight photoreceptor (R) and the five secondary pigment (SP) cells of the adult ommatidium2,3. These cells lie close together within the ommatidium and are easily distinguishable with morphological criteria. We asked the following question: do R and SP cells have common precursors at the latest developmental stage or do they belong to different cell lineages?

We have analysed the effect of temperature on both developing and adult eye cell clones homozygous forshi

Like horizontal cells in vertebrate retinas, horizontal amacrine cells beneath the insect eye intervene between receptors and interneurons at the first level of synapses. Synaptic arrangements between amacrines and interneurons that give rise to regular networks of axon collaterals may explain recent electrophysiological observations of lateral inhibition beneath the insect retina. Neural adaptation mechanisms acting on single retinotopic channels or assemblies of channels can also be referred to reciprocal relationships between receptors and first-order interneurons as well as to centrifugal cells from levels of so-called photopic receptor endings.

Complex biological processes, such as the development of the central nervous system (CNS) of Drosophila, can be dissected into logical steps by introducing genetic variations. In the beginning the major difficulty encountered in this process is that of defining the elements involved, which in turn is dependent on obtaining mutants suitable for analysis. The reason for this difficult is twofold: on the one hand, mutants affecting essential aspects of neurogenesis are expected to be lethal; on the other hand, the CNS is an internal structure, and the appraisal of mutant variations requires the use of special techniques. Drosophila goes through different developmental stages, embryonic, larval and pupal, before reaching the final imaginal stage. Assuming a low probability for neurogenic mutants to develop to the imaginal stage, a reasonable way of screening for such mutants is to concentrate on the study of embryos. For neurobiological studies Drosophila embryos are inconvenient because of their small size, which makes selective staining of neurones extremely difficult. However, they offer several important advantages for our purposes: the CNS of mature embryos is roughly comparable to that of third instar larvae, disregarding the proliferation within the imaginal anlagen; techniques have been developed which allow the simultaneous study of dozens of embryos as whole-mounts, in which the CNS can be visualized “in toto”; embryonic lethality can be used as a previous selection criterium. Wright (1970) proposed selecting for embryonic lethal point mutants when searching for mutants affecting embryogenesis; the same strategy could be pursued for studying neurogenic variations since they are probably found among embryonic lethals. However, a few arguments make clear that this approach may be very time consuming. Mutants with embryonic pheno-effective periods have been found in 30% of all point lethals (Hadorn, 1961); furthermore, mutagenesis studies have demonstrated that about 90% of all genes can mutate to lethal conditions (Judd et al., 1972). Accepting that the total number of genes in the chromosomal genome of Drosophila is about 5000 (Lefevre, 1974) a simple calculation indicates that about 1350 genes might show an embryonic lethal phenotype when mutated.

The uptake of 3H-GABA in the visual system of half-head preparations of Musca and Drosophila was studied by means of light and electron microscope autoradiography. Of all three ganglia, only the first synaptic region, the lamina ganglionaris, showed accumulation of radioactive grains, and there a preferential glial uptake could be found. Under normal light conditions at incubation (constant light flux of 100 Lux) the maximum of radio-activity was found in the marginal glia cells. Increasing the time of incubation produced also an increase in the number of grains per surface unit in the marginal glia cells. After changing the light intensity during incubation, quantitative modifications of the distribution of radio-activity were observed: incubating with stroboscopic illumination, the number of grains diminished in the marginal glia cells and remained constant in the epithelial cells; incubated in darkness, the epithelial cells became more intensely labelled whilst the number of grains decreased in the marginal cells. The possibility is discussed that the receptor axons 1–6 are the neurological elements of the lamina which use GABA as a transmitter. This hypothesis is lent some support from results of similar experiments with neurological mutants of Drosophila. In opm 18 there was a delayed uptake of 3H-GABA whereas in opm 3 and ebony the results were comparable to those found in Musca incubated in darkness. Behavioral studies on these mutants have demonstrated a defect, most probably related to the receptor system 1–6.

The projections of long visual fibres (R7+R8) from the retina to the medulla and their levels of termination in the second synaptic region have been studied with the aid of light and electron microscopy. Normal and degenerated material has revealed the fine structural characteristics of the long visual fibres as well as their orientations relative to those of a pair of first order interneuron terminals (monopolar cells). R7 and R8 project as pairs across the first optic chiasma. Their position equatorial to monopolar cells in the lamina is shifted to a polar position in the medulla via the chiasma. Their endings are situated at two levels: namely, R7 is deeper than R8. Like the short visual fibre endings in the lamina (R1–R6) the terminals of R7 and R8 in the medulla have synaptic specialisations throughout their length. The long visual fibres provide a point for point projection from the retina to the medulla and with monopolar cell terminals are arranged coincident with the three axes of the visual neuropil.

1. The cellular and synaptic pattern of the lateral geniculate body of the monkey has been studied electronmicroscopically by using microdissected portions from individual layers. For general orientation and comparison the Golgi technique has been utilised extensively. 2. This technique revealed four different types of neuron whose distribution differs in the various layers. The first type has 3–4 stem dendrites, which possess tiny side branches or appendages. The second type is smaller and has 8–10 stem dendrites. The third type is a Golgi type II cell with very long dendrites with bushy terminal arborisations. The fourth type are very small neurons and it is difficult to separate them from glial cells with certainty. 3. The synaptic organisation has been examined in great detail, in particular the different composition of the glomeruli in different layers. Seven components of a glomerulus have been found whose individual contributions vary in different layers. These components are: main axon, main dendrite, peripheral axons, peripheral dendrites, dendritic and axonal thorns and glia lamellae. 4. In addition to this synaptic pattern a further interesting synapse has been found, an axo-axonal type in contact with the initial segment of a large axon of a geniculate cell.

Combination of Golgi staining, reduced silver staining, normal and Golgi-E.M. has allowed estimates about the numbers of cells present in medullary columns. Up to the present time 120 cell types, classified according to orientation, depth, dendritic spread and projections, have been identified. Components of 34 cell types occur in each medullary column. In addition, each column contains the corresponding medullar components of elements destined for or derived from an optic cartridge, and the R7 / R8 endings. The basic three dimensional structure of the medulla, as defined by the 46 cells that contribute to each column, is columnar and stratified. Information flow must be both lateral and from periphery to centre or vice versa. This is in contrast with the lamina structure where most information flow must be along a cartridge. This structure is discussed with respect to the remaining cell types whose distribution per column is still unknown.

The first synaptic region of the fly's visual system contains sets of climbing fibres enclosing each optic cartridge in a double basket-like arrangement of processes. The two sets of processes, termed α and β, can be separately identified by their fine structure. αfibres are more electron dense than β; they are postsynaptic to receptor endings from the retina and contain synaptic specializations apposed to epithelial cell membranes, to β fibres and to the distal processes of the L4 monopolar neurone. β is postsynaptic to receptors and to α. It has a lighter cytoplasm and contains peculiar glial invaginations called ‘gnarls’. Degeneration experiments, where the optic chiasma between the 1st and 2nd synaptic regions has been severed, demonstrate that α fibres are intrinsic to the lamina. Golgi impregnation and electron microscopy of Golgi-impregnated cells have demonstrated that β fibres belong to a type of medulla-to-lamina cell, T1, whereas α fibres belong to an intrinsic (amacrine) cell in the lamina. The amacrine cells have wide fields through several cartridges. However, Golgi—EM studies show that each cartridge must be invaded by the processes of at least two cells. The synaptology of these elements is discussed with respect to previous data about the lamina, and certain analogies with the vertebrate plexiform layer are drawn.

Electron microscopy of Golgi impregnated neurons has allowed the precise determination of the spatial relationships between identifiable profiles in optic cartridges and corresponding terminals in the medulla columns. Golgi-E.M. profiles can be correlated with profiles in normal electron microscopy. Structural contacts of cells in the lamina as well as their corresponding components in the medulla are described: functional contacts (synapses) known up to the present time are listed in figure 5 (polarities arrowed). Possibly structural contact may be indicative of functional intimacy, but this is still to be determined. At least three centripetal pathways (L1, L2, L3) convay information from R1 — R6 into the medulla. The other monopolars L4 and L5 appear to be intimate with endings of centrifugal cells from the medulla. The logic underlying the complexity of lamina organization seems to lie in the arrangement of feed-back and feed forward loops in parallel with the centripetal cells.

After experimental lesions inGalago crassicaudatus, the cortico-efferent fibres of the optical cortex were studied using theNauta andGlees methods. Mainly 4 different systems were separated: Corticocaudate, pontine, geniculate and pulvinar.