Embryonic and postembryonic neuroblasts in the thoracic ventral nerve cord of Drosophila melanogaster have the same origin. We have traced the development of threefold-labelled single precursor cells from the early gastrula stage to late larval stages. The technique allows in the same individual monitoring of progeny cells at embryonic stages (in vivo) and differentially staining embryonic and postembryonic progeny within the resulting neural clone at late postembryonic stages. The analysis reveals that postembryonic cells always appear together with embryonic cells in one clone. Further-more, BrdU labelling suggests that the embryonic neuroblast itself rather than one of its progeny resumes proliferation as a postembryonic neuroblast. A second type of clone consists of embryonic progeny only.

In hemimetabolous insects, the CNS of the adult is virtually established during embryogenesis (Bate, 1976; Doe and Goodman, 1985; Malzacher, 1968). In holometabolous insects, however, there are two discrete periods of neurogenesis.

During embryogenesis, a set of neuroblasts (NBs) delaminates from the ectodermal layer producing the neural cells that make up the CNS of the larva (Wheeler, 1891; Poulson, 1950). In the Drosophila embryo, about 500 NBs delaminate from the ventral neurogenic region (vNR; Hartenstein and Campos-Ortega, 1984). They become arranged in repeating patterns of about 40 NBs per segment (Hartenstein et al. 1987). With each division, these embryonic stem cells decrease in volume and they are no longer identifiable as NBs in the late embryo (Hartenstein and Campos-Ortega, 1984; Hartenstein et al. 1987). Their fate is unknown.

Metamorphosis of the larva into the adult requires a drastic remodelling of the nervous system (Hertweck, 1931; Weismann, 1864). A large number of ganglion cells is added during postembryonic development (Bauer, 1904; Nordlander and Edwards, 1969; White and Kankel, 1978; Booker and Truman, 1987; Truman and Bate, 1988) and will – in addition to persisting embryonic cells – build up the imaginai nervous system. At hatching, a larval ventral neuromere of Drosophila contains about 300 cells (Poulson, 1950). About 4000 new ganglion cells are added to each thoracic neuro-mere during the larval period by the proliferative activity of postembryonic NBs. The first of these NBs can be detected in the first instar larva as they increase in volume and, subsequently, commence proliferation. As in the embryo, they appear in repeating patterns containing about 47 postembryonic NBs in each thoracic segment (Truman and Bate, 1988). Genetic mosaics for enzyme markers revealed that these postembryonic NBs map to the same region of the blastodermal fate map as the embryonic stem cells (Kankel and Hall, 1976). Several authors have discussed the possibility that embryonic and postembryonic NBs may be the same cells (White and Kankel, 1978; Truman and Bate, 1988), but the origin of postembryonic NBs is still unknown and so are the mechanisms leading to the spatial distribution of these stem cells in the larval ventral nerve cord (vNC).

Here we describe a method that traces the development of single precursor cells from the early embryo to late postembryonic stages and discriminates between embryonic and postembryonic progeny cells. We show for thoracic ganglia that postembryonic cells are descended from the same precursors as embryonic cells, as both appear in the same clone. Furthermore, 5-bromodeoxyuridine (BrdU) incorporation into presumptive postembryonic NBs at late embryonic stages suggests that these cells are identical to the embryonic NBs.

Stocks

Transformant strains carrying a P-element vector with the P–lacZ fusion gene were used as donors for cell transplantation: A45/CyO (courtesy of C. O’Kane; O’Kane and Gehring, 1987), E and Bl (insertion on the first and second chromosome respectively; courtesy of Y.-N. Jan; Bier et al. 1989). As hosts we used the β -Gal-1 strain cq20 lacking endogenous β-galactosidase activity (courtesy of R. J. MacIntyre; Knipple and McIntyre, 1984). BrdU was injected into wild-type embryos (Oregon R).

Cell labelling and transplantation

Staged embryos were dechorionated with a needle, stuck to a cover-slip, dried and covered with fluorocarbonoil. 5-10 nl of a solution of 10% horseradish peroxidase (HRP) and 5% rhodamine-isothiocyanate-dextran (RITC-dextran) in 0.2 M KC1 were injected into donors at the syncytial blastoderm stage. The label becomes incorporated in all cells at cell formation (Technau, 1986). In addition, the donor embryos carried the P-ZacZ fusion gene expressing /3-galactosidase in all neural cells. At transplantation, donor and host embryos were of the same age, approximately 10 min after the onset of gastrulation (stage 7; all embryonic stages mentioned refer to Campos-Ortega and Hartenstein, 1985). Up to 10 cells were removed from the thoracic ventral neurogenic region (vNR) of donors and transplanted individually into the vNR of unlabelled hosts.

Single cell implantation into hosts was inspected with Nomarski optics during transplantation and, in more than 30 % of the cases, they were scrutinized with the fluorescence microscope five minutes later. Cases of transplantation of two or more cells were discarded.

Staining and analysis of neural clones

Fluorescence-labelled clones in living host embryos were documented on a video tape a few hours before hatching (stage 17). Soon after hatching larvae were individually transferred to vials containing about 1 ml diet and allowed to continue development until they became wandering larvae or adults. The central nervous system (CNS) was removed and, after fixation in 1 % glutaraldehyde (in 0.2 M phosphate buffer (PB), pH7.2), stained for β-galactosidase activity in X-Gal solution (adding one part of 20 % X-Gal in DMF to 99 parts of a warm (60°C) solution of 150HIM NaCl, 1 HIM MgC12, 3.3 mM K3[Fe(CN)6] in PB) overnight at room temperature. The preparations were screened for labelled cells under a dissecting microscope (magnification: 50x). Afterwards, ft-Gal-positive individuals were separated from negative ones. Both were stained for HRP with diaminobenzidine (0.1% DAB, 0.02% H2O2 in PB) for about 10 min at room temperature. After removal of the hemispheres and dehydration, the ventral nerve cord was transferred to DePeX (Serva) and sucked into a borosilicate capillary (inner diameter 0.2 mm) allowing microscopic inspection of the whole mounts from all sides.

Injection and staining of BrdU

Staged embryos were prepared for injection as described previously. A 15 mM solution of BrdU (Sigma) in 0.2 M KC1 was injected (see also Shepherd and Bate, 1990) at defined stages of embryogenesis. The embryos were allowed to develop until hatching. They were staged a second time by selecting the hatched larvae every 90min (at 25°C). The larvae were fed on standard medium for 28 or 34 h, the CNS was removed and stained for BrdU incorporation according to Truman and Bate (1988). A second sample of embryos at parallel stages were sham-injected with 0.2 M KC1 and from the time of hatching were fed on a diet containing 0.1 % of a BrdU solution (33 mM BrdU in 40% ethanol; according to Truman and Bate, 1988). Preparation and staining of their CNSs was the same as for the first sample. BrdU preparations were embedded in Araldite (Serva).

There are two principal possibilities with regard to the origin of postembryonic NBs. First of all, they may derive from a population of cells totally distinct from the embryonic NBs. In this case, one would assume that among the cells delaminating from the neurogenic ectoderm, a subpopulation is kept mitotically silent during embryogenesis in order to become active as postembryonic NBs in the larva (see A in Fig. 1). The second possibility is that postembryonic NBs have the same origin as embryonic NBs (see B in Fig. 1).

Fig. 1.

Schematic presentation of three possible kinds of relationship between embryonic (E) and postembryonic (P) NBs. Following delamination from the neurogenic ectoderm, NBs bud off a chain of smaller ganglion mother cells which divide once more to give rise to two neurones (Bauer, 1904; Bate, 1976; Hartenstein et al. 1987). Either both types of NBs represent different cell populations already in the blastoderm (A) or they have a common origin (B). In the latter case, the lineage could either bifurcate (Bl) or the embryonic NB itself resumes proliferation during postembryonic development (B2). In case of Bl, the postembryonic NB would develop from one of the progenies of the embryonic NB whereas the embryonic NB would be used up by its last division in the embryo or degenerate. In case of B2, embryonic and postembryonic NB would be identical.

Fig. 1.

Schematic presentation of three possible kinds of relationship between embryonic (E) and postembryonic (P) NBs. Following delamination from the neurogenic ectoderm, NBs bud off a chain of smaller ganglion mother cells which divide once more to give rise to two neurones (Bauer, 1904; Bate, 1976; Hartenstein et al. 1987). Either both types of NBs represent different cell populations already in the blastoderm (A) or they have a common origin (B). In the latter case, the lineage could either bifurcate (Bl) or the embryonic NB itself resumes proliferation during postembryonic development (B2). In case of Bl, the postembryonic NB would develop from one of the progenies of the embryonic NB whereas the embryonic NB would be used up by its last division in the embryo or degenerate. In case of B2, embryonic and postembryonic NB would be identical.

A method for differentially labelling embryonic and postembryonic cells within individual CNS clones

In order to find out whether or not the two types of NBs are clonally related, we established a method that allows us to label differentially the embryonic and postembryonic progeny of individual blastoderm cells (Fig-2).

As a genetic marker we used the enzyme β-galactosidase of E. coli. Different transformant strains that carried a P-transposon with lacZ as a reporter gene (O’Kane and Gehring, 1987; Bellen et al. 1989; Bier et al. 1989) were screened for strong expression of lacZ, especially in the CNS of wandering larvae. The strain A45 is homozygous lethal but expresses lacZ very strongly. The strains Bl and E are homozygous viable but not as strong in expression. As their insertions were located on different chromosomes, we combined them by crossing and established the strain E;B1 with four copies of lacZ. We inspected all strains (A45, Bl, E and B1;E) for β-galactosidase expression in the CNS of third instar larvae and there was no indication of CNS cells that did not express the enzyme (data not shown). This was confirmed by staining with an anti-β-Gal antibody. By using these transformant strains as donors for transplantation of early gastrula cells into host embryos without any endogenous β-galactosidase activity (cq20; Knipple and McIntyre, 1984), all of their progeny cells become distinctly labelled even in late larval or adult stages (Fig. 3). As each cell produces galactosidase, there is no dilution of the marker.

Fig. 2.

Schematic presentation of the method used for tracing the development of single blastodermal precursor cells through embryonic and postembryonic stages and differentially labelling embryonic and postembryonic progeny cells. (A) A mixture of HRP and RITC-dextran is injected into donors during the syncytial blastodermal stage or during cellularization and diffuses throughout the embryo. The strain that we use as donors is genetically labelled and expresses β-Gal in all CNS cells. (B) After cell closure the injected markers become incorporated into all blastodermal cells. Some of these threefold labelled cells of the neurogenic ectoderm are sucked into a capillary at stage 7 (about 10 min after the onset of gastrulation) and (C) are singly transplanted into unlabelled host embryos of the same age. (D) Clones originating from transplanted cells can be monitored in vivo under the fluorescence microscope and documented on a video tape during late embryogenesis (see close-up). (E) The hatching larvae are individually raised until they start wandering, the CNS is removed and stained for β-Gal and subsequently for HRP (close-up showing a labelled clone). (F) Composition of clones can also be analyzed in the imaginai stage. Bar, 20 μm.

Fig. 2.

Schematic presentation of the method used for tracing the development of single blastodermal precursor cells through embryonic and postembryonic stages and differentially labelling embryonic and postembryonic progeny cells. (A) A mixture of HRP and RITC-dextran is injected into donors during the syncytial blastodermal stage or during cellularization and diffuses throughout the embryo. The strain that we use as donors is genetically labelled and expresses β-Gal in all CNS cells. (B) After cell closure the injected markers become incorporated into all blastodermal cells. Some of these threefold labelled cells of the neurogenic ectoderm are sucked into a capillary at stage 7 (about 10 min after the onset of gastrulation) and (C) are singly transplanted into unlabelled host embryos of the same age. (D) Clones originating from transplanted cells can be monitored in vivo under the fluorescence microscope and documented on a video tape during late embryogenesis (see close-up). (E) The hatching larvae are individually raised until they start wandering, the CNS is removed and stained for β-Gal and subsequently for HRP (close-up showing a labelled clone). (F) Composition of clones can also be analyzed in the imaginai stage. Bar, 20 μm.

Fig. 3.

Examples of clones in thoracic neuromers labelled with β-Gal and HRP. (A) Neural clone in the late larva. Embryonically derived neurones in the inner cortex (Cx) stain for both cell markers. They send projections into the neuropile (N). The postembryonic NB (arrowhead) lies in the periphery associated with a nest of postembryonic progeny. In these cells the HRP is diluted out. (B) Neural clone in the imaginai nerve cord. The blue postembryonically derived cells are no longer arranged in a single compact nest. Cells labelled with HRP are of embryonic origin and persisted through metamorphosis (small arrows). The large cells probably are glial cells (big arrows); their cytoplasm is also labelled with HRP. (C – F) Some other examples of neural clones in the larval ventral cord. They all contain embryonically as well as postembryonically derived cells. Arrowheads in C, D and F point to the postembryonic NB The corresponding in vivo documentation of the same clones in the late embryo is presented in Fig. 4C – F. Bar, 20 μm.

Fig. 3.

Examples of clones in thoracic neuromers labelled with β-Gal and HRP. (A) Neural clone in the late larva. Embryonically derived neurones in the inner cortex (Cx) stain for both cell markers. They send projections into the neuropile (N). The postembryonic NB (arrowhead) lies in the periphery associated with a nest of postembryonic progeny. In these cells the HRP is diluted out. (B) Neural clone in the imaginai nerve cord. The blue postembryonically derived cells are no longer arranged in a single compact nest. Cells labelled with HRP are of embryonic origin and persisted through metamorphosis (small arrows). The large cells probably are glial cells (big arrows); their cytoplasm is also labelled with HRP. (C – F) Some other examples of neural clones in the larval ventral cord. They all contain embryonically as well as postembryonically derived cells. Arrowheads in C, D and F point to the postembryonic NB The corresponding in vivo documentation of the same clones in the late embryo is presented in Fig. 4C – F. Bar, 20 μm.

As a second marker, we injected HRP into syncytial blastoderm donor embryos (of transformant strains). This enzyme is still active in adults but it becomes almost completely diluted-out in postembryonic NBs of the vNC, when they increase in volume during early larval stages. Thus, in postembryonic NBs and their progenies, staining for HRP is almost negative. Progeny cells, however, which originate from embryonic NBs remain intensely labelled. Thus, injection of HRP provides a means of distinguishing between cells of embryonic and postembryonic origin (Tix et al. 1989a, b).

As a third marker, we coinjected RITC-dextran. With the help of this marker it is possible with a fluorescent microscope to ensure shortly after the transplantation that’ only a single cell has been transplanted. Furthermore, the fluorescent label provides a means of checking in vivo at late embryonic stages whether or not the transplanted precursor cell had produced progeny cells during embryogenesis. This was documented on a video tape (Fig. 4).

Fig. 4.

(A,C – F) Fluorescence labelled neural clones in the thoracic vNC of living stage 17 embryos. Broken lines outline the vNC. (B) Camera lucida drawing of the clone at the late third instar larval stage of the same individual shown in A. Size and position of the fluorescence-labelled clone in A correspond to the HRP-labelled subclone (dark cells) in the vNC of the larva (B). These embryonically derived cells are fully differentiated, carrying fibers that project along the ipsi- and contralateral connectives in an anterior and posterior direction. The commissural fascicle (small arrow in B) can be seen in the embryo already (arrow in A). Light cells in B represent the postembryonic NB (arrowhead) and its progenies. Broken lines (in B) delimit the cortex/neuropil border in the vNC. Circles along the midline (big arrows) mark the position of ‘channels’ crossing the vNC ventrodorsally near the segment borders. C – F show the same individuals as in Fig. 3C – F. Fluorescently labelled clones correspond to the HRP-labelled subclones in Fig. 3. Cx, cellular cortex; N, neuropile; Bars, 20 μm.

Fig. 4.

(A,C – F) Fluorescence labelled neural clones in the thoracic vNC of living stage 17 embryos. Broken lines outline the vNC. (B) Camera lucida drawing of the clone at the late third instar larval stage of the same individual shown in A. Size and position of the fluorescence-labelled clone in A correspond to the HRP-labelled subclone (dark cells) in the vNC of the larva (B). These embryonically derived cells are fully differentiated, carrying fibers that project along the ipsi- and contralateral connectives in an anterior and posterior direction. The commissural fascicle (small arrow in B) can be seen in the embryo already (arrow in A). Light cells in B represent the postembryonic NB (arrowhead) and its progenies. Broken lines (in B) delimit the cortex/neuropil border in the vNC. Circles along the midline (big arrows) mark the position of ‘channels’ crossing the vNC ventrodorsally near the segment borders. C – F show the same individuals as in Fig. 3C – F. Fluorescently labelled clones correspond to the HRP-labelled subclones in Fig. 3. Cx, cellular cortex; N, neuropile; Bars, 20 μm.

Hatched larvae were allowed to develop to the late third larval stage, when their CNS was removed and histochemically stained for β-galactosidase and, subsequently, for HRP activity.

Thus, the technique consists of the transplantation of a threefold-labelled precursor cell into an unlabelled host approximately 10 min after the onset of gastrulation, with subsequent in vivo inspection of the clone during embryonic development and differential staining for embryonic and postembryonic progeny within the same clone at postembryonic stages.

Composition of clones

We found two types of clone in the vNC of late third instar larvae. The first type of clone consisted of HRP and β-galactosidase double-labelled cells only. The second type contained double-labelled cells and, in addition, cells only labelled with β-galactosidase.

The second type of clone (Fig. 3) comprises one large circular cell (7 – 11 μm in diameter) that is located at the periphery of the cortex, either immediately beneath the neuronal sheath or up to three cell layers deeper. As indicated by its size and position, this cell is a postembryonic NB (see also Truman and Bate, 1988),. It appears indigo coloured as it expresses the lacZ gene. Its nuclear membrane and cell membrane are distinctly stained. Typical for this type of clone is a nest of small blue cells. Their number seems to vary from clone to clone. Clones of late third instar larvae may comprise up to 200 cells. Because of their small size (3–4 μm) and their arrangement as a compact cluster, we were unable to determine their exact number. The nest lies immediately beneath the postembryonic NB. From the margin of the cortex the nest extends inwards where it contacts the HRP-labelled cells. Although massive proliferation during larval life leads to a rearrangement of spatial relationships in the thoracic cortex of the vNC (Truman and Bate, 1988), the blue staining nest of cells does not normally lose contact with the HRP-labelled cells. The HRP-labelled cells are larger (4 – 6 μm in diameter) and are concentrated in the inner layers of the cortex near the neuropile. Their number ranges from one to about 25 cells. They also express lacZ, which is particularly evident at the nuclear membrane (Fig. 3A). The brown HRP reaction product, however, is more intense and stains the entire cell. HRP-labelled neurones as .well as glial cells are fully differentiated (Fig. 4B). Neuronal projections carry spines and glial processes can be seen to ensheath several structures of the vNC (not shown).

The other type of clone, containing exclusively HRP and β-galactosidase double-labelled cells, was not essentially different in its morphological appearance from the HRP-labelled parts of those clones, which, in addition, carry a nest of purely blue cells. No NB is stained in this type of clone.

Identity of differentially labelled subclones

The following data suggest that the HRP-labelled subclones consist of cells that originate during embryogenesis: they are located in the inner cortex region. Their relatively small sizes (up to 25 cells) correspond to those of CNS clones found in the late embryo (Technau and Campos-Ortega, 1986). Finally, these cells are fully differentiated as neurones or glial cells. In order to confirm the embryonic origin of these cells, we documented the clones in vivo (by video fluorescence microscopy) at late embryonic stages, when all larval cells have been produced. In all 21 cases, the fluorescent clone recorded in the late embryo corresponded in shape and size to the HRP-labelled clone or subclone later found in the third instar larva, thus proving its embryonic origin.

The purely β-galactosidase-labelled subclones, on the other hand, are located in the peripheral cortex in close contact with the postembryonic NBs. This and the fact that they comprise a large number of cells that do not stain for HRP point to their descent from the postembryonic NB.

From these results, we conclude that there are two types of clones in the developing imaginai CNS. One type consists exclusively of cells that originate during embryogenesis from embryonic NBs. The other type of clone consists of two subclones. One subclone also consists of embryonically derived cells, the other subclone corresponds to a nest of cells that originates during postembryonic development from the postembryonic NB. Both subclones are derived from the same blastodermal cell. Thus, postembryonic NBs always share common lineages with embryonic NBs. Possibility A in Fig. 1 has to be rejected.

Frequency of the two types of clones

Homotopic transplantations were carried out in the thoracic vNR at 55 – 60% egg length (EL; 0% EL posterior pole) of the early stage 7 embryo. We obtained 66 thoracic clones carrying the genetic marker in the late larval CNS (Table 1).

Table 1.

Frequencies of two types of clones obtained in the thoracic ventral cord of third instar larvae following transplantation of single precursor cells at the early gastrula stage

Frequencies of two types of clones obtained in the thoracic ventral cord of third instar larvae following transplantation of single precursor cells at the early gastrula stage
Frequencies of two types of clones obtained in the thoracic ventral cord of third instar larvae following transplantation of single precursor cells at the early gastrula stage

45 (68 %) of these clones contained an embryonic and a postembryonic subclone. In 38 (84%) out of these clones, we identified one postembryonic NB. In 7 preparations (16%), no NB was visible, probably because it becomes obscured by the intensely stained nest of postembryonic progeny cells. There were two preparations containing 2 postembryonic NBs. Each of these NBs was associated with a separate cluster of cells including embryonic and postembryonic components. Thus, these preparations appeared to contain two clones.

21 (32%) clones were composed of embryonically derived cells only. These were produced by cells from both transformant strains used (A45 and E;B1; Table 1). Clones without postembryonic cells are not spatially restricted, appearing in all ventrodorsal positions of the thoracic vNC at similar frequencies (data not shown).

Distribution of postembryonic NBs in the blastoderm fate map

Transplantations were carried out in stage 7 embryos approximately 10 min after the onset of gastrulation. According to the larval fate map of that stage (Technau and Campos-Ortega, 1986; Hartenstein et al. 1985), cells were placed into the vNR between 0 % and 50 % of the ventrodorsal dimension (VD; 0% VD ventral furrow) within the anlagen of the thoracic segments. Postembryonic NBs of the resulting neuronal clones came to lie in all ventrodorsal positions of the vNC where they map under normal conditions (Truman and Bate, 1988). The most ventral NBs were lying near the ventral midline of the larval vNC. Two of these cases originate from transplantations carried out homotopically at about 0% VD. The most dorsal NBs were lying between 40% and 90% dorsoventral circumference of the larval vNC. Four of these cases originate from transplantations carried out homotopically at about 50% VD. From these results, we conclude that postembryonic NBs map in the same region of the blastodermal fate map as the embryonic NBs, which agrees with the fact that they emerge from the same precursor cells.

Are embryonic and postembryonic NBs identical?

While our clones demonstrate that postembryonic NBs stem from the same precursors as embryonic NBs they do not tell us whether the embryonic NB itself (see Bl in Fig. 1) or one of its daughter cells (see B2 in Fig. 1) will resume proliferation during postembryonic development.

To test this, we injected BrdU into staged embryos of increasing age. This substituted nucleoside becomes incorporated during DNA synthesis and can be detected immunocytochemically with a monoclonal antibody against BrdU (Gratzner, 1982). If embryonic NBs and postembryonic NBs are identical, postembryonic NBs should become labelled whenever BrdU is injected before the last DNA synthesis in embryonic NBs. If, on the other hand, the postembryonic NB is set aside at an earlier stage, it should incorporate the tracer only until the point when it becomes separate from the embryonic NB, given that DNA synthesis is restricted to premitotic cells. Injections were carried out at stage 12, at late stage 15, at stage 16 and during stage 17 (Fig. 5). In all cases, hatched larvae were allowed to develop further for 28 to 34 h at 25 °C, when their CNS was removed and stained for BrdU. At this stage, the first postembryonic NBs start DNA synthesis and subsequently proliferation: in 28 h old larvae which were sham-injected with 0.2 M KC1 during embryogenesis and exposed to a BrdU-containing diet from the time of hatching, the tracer became incorporated into postembryonic NBs some of which were associated with labelled progeny cells (Fig. 6E; see also Truman and Bate, 1988). Many of the postembryonic NBs have not yet attained their final size and thus are not identifiable by morphological criteria.

Fig. 5.

Diagram showing the times when BrdU was injected into embryos. The scale on the left side indicates the embryonic stages according to Campos-Ortega and Hartenstein (1985). Scale on the right side represents the time (h) after fertilization at 25°C. Stars indicate the eight parasynchronous mitotic cycles of embryonic NBs described by Hartenstein et al. (1987). Arrows indicate the times when injections of BrdU were carried out. A – D correspond to A – D in Fig. 6.

Fig. 5.

Diagram showing the times when BrdU was injected into embryos. The scale on the left side indicates the embryonic stages according to Campos-Ortega and Hartenstein (1985). Scale on the right side represents the time (h) after fertilization at 25°C. Stars indicate the eight parasynchronous mitotic cycles of embryonic NBs described by Hartenstein et al. (1987). Arrows indicate the times when injections of BrdU were carried out. A – D correspond to A – D in Fig. 6.

Fig. 6.

Labelled cells in the CNS of 28 to 34 h old larvae following injections of BrdU at delined embryonic stages. From top to bottom the pictures show a dorsal view of the hemispheres, a ventral view of the vNCs and a close-up of thoracic neuromeres. A – D show the pattern of BrdU-labelled cells (except close-up in D) following injections during embryogenesis as indicated in Fig. 5. In the vNC, injections at stages 12 (A), 15 (B) and 16 (C) lead to incorporation of BrdU, and the number of labelled cells decreases from A to C. In all these cases, postembryonic NBs (see arrowheads in close-ups) are labelled together with smaller cells. (D) Following injections during stage 17 no more incorporation can be seen in the vNC except in a few single cells. Although postembryonic NBs in larvae of the same experiment are clearly visible upon staining with toluidine blue (according to Truman and Bate, 1988; see bottom of D) none of them is labelled with BrdU. In the brain hemispheres (see top row), on the other hand, many cells still incorporate BrdU upon injection at stage 17. E presents the CNS of a 28 hour old larva that was sham-injected with 0.2 M KC1 during embryogenesis and transferred to BrdU-containing diet from the time of hatching. Especially in the lateral region (arrows) postembryonic NBs incorporated BrdU. Several of them have already started proliferation as they are associated with labelled progeny cells. Bars, 20 μm.

Fig. 6.

Labelled cells in the CNS of 28 to 34 h old larvae following injections of BrdU at delined embryonic stages. From top to bottom the pictures show a dorsal view of the hemispheres, a ventral view of the vNCs and a close-up of thoracic neuromeres. A – D show the pattern of BrdU-labelled cells (except close-up in D) following injections during embryogenesis as indicated in Fig. 5. In the vNC, injections at stages 12 (A), 15 (B) and 16 (C) lead to incorporation of BrdU, and the number of labelled cells decreases from A to C. In all these cases, postembryonic NBs (see arrowheads in close-ups) are labelled together with smaller cells. (D) Following injections during stage 17 no more incorporation can be seen in the vNC except in a few single cells. Although postembryonic NBs in larvae of the same experiment are clearly visible upon staining with toluidine blue (according to Truman and Bate, 1988; see bottom of D) none of them is labelled with BrdU. In the brain hemispheres (see top row), on the other hand, many cells still incorporate BrdU upon injection at stage 17. E presents the CNS of a 28 hour old larva that was sham-injected with 0.2 M KC1 during embryogenesis and transferred to BrdU-containing diet from the time of hatching. Especially in the lateral region (arrows) postembryonic NBs incorporated BrdU. Several of them have already started proliferation as they are associated with labelled progeny cells. Bars, 20 μm.

Following injection at embryonic stage 12 (Fig. 6A), many BrdU-labelled postembryonic NBs were visible in the thoracic neuromeres, among them the four midline NBs (Truman and Bate, 1988). Their nuclei are about twice as big as those of the surrounding darker staining ganglion cells. The staining of ganglion cells appeared to be quite homogenous although somewhat reduced in the abdomen. Those parts of the cortex, however, lying dorsal to the neuropile did not stain. Incorporation of BrdU into cells of this region is only achieved, after injection during earlier embryonic stages (e.g. stages 7 or 10).

The CNS of larvae that received BrdU at late stage 15 (Fig. 6B) showed a smaller number (an average of 46 cells per hemineuromere) but a similar distribution of labelled cells. In the thorax, postembryonic NBs that had incorporated BrdU were intermingled with unlabelled NBs. The midline NBs were no longer labelled.

After injection at stage 16 (Fig. 6C), there was still incorporation of BrdU although in a smaller number of cells, mainly restricted to the thorax. Thus, completion of embryonic DNA synthesis in the abdomen seems to precede that in the thorax. In the lateral cortex of each thoracic hemineuromere, we found 4 – 6 NBs labelled with BrdU while in the ventral cortex only one postembryonic NB was labelled. The labelled NBs were associated with two or sometimes three small labelled progeny cells. In the abdomen single or pairs of small labelled cells occurred in the lateral cortex. This suggests that, in this series, injections preceded the last embryonic division of these cells.

Injections carried out at stage 17 (Fig. 6D) did not lead to any incorporation of BrdU in the vNC with the exception of one or two single cells in some of the preparations. Therefore, in the vNC all embryonic NBs have stopped dividing by embryonic stage 17. In the brain hemispheres, however, there are still some groups of labelled cells in characteristic positions.

In order to see whether postembryonic mitoses of NBs would still lead to incorporation of BrdU, we allowed some of the individuals injected during embryonic stage 17 to develop until they became wandering larvae. These preparations also did not show any BrdU incorporation in the vNC (data not shown). Thus, in our experiments incorporation of BrdU into postembryonic NBs is due to synthetic activity of these cells during embryogenesis. The fact that some postembryonic NBs stain positively following BrdU injection at late stage 16 indicates that these postembryonic NBs emerge from the cells that are the last to perform DNA synthesis in the embryonic CNS. This suggests that postembryonic NBs are identical with embryonic NBs or one of their latest embryonic progeny cells (see B2 in Fig. 1).

Common origin of embryonic and postembryonic neuroblasts

We have developed a method that allows us to trace the fate of single blastodermal precursor cells through embryonic and postembryonic development and to label differentially their embryonic and postembryonic progeny cells. The method is based on threefold labelling and transplantation into unlabelled hosts of single precursor cells from the neurogenic ectoderm. This method has allowed us to show that postembryonic NBs have the same origin as embryonic NBs.

A prerequisite for this method is that it should provide good control over single transplanted cells. Single cell transplantation can be monitored by direct observation with Nomarski optics. In those cases where we used RITC-dextran as an additional marker, we scrutinized the single cell transplant by fluorescence microscopy. In two of the late larval CNSs we found two cell clusters, each consisting of embryonic and postembryonic neurones and a postembryonic NB. We cannot exclude that in these cases transplantation of two precursor cells may have occurred, since these transplants were not confirmed under the fluorescence microscope. On the other hand, it could well be that the transplanted cell performed its first division whilst still in the ectoderm and that both daughter cells delaminated as NBs to give rise to two separate neural subclones (as also discussed in Technau and Campos-Ortega, 1986; Technau, 1987).

A second prerequisite for the applicability of our approach is a ubiquitous expression of /3-galactosidase in the CNS of the donor strains, especially in late third instar larvae. X-Gal staining of both donor strains as well as anti-/8-Gal antibody staining of A45 larvae revealed strong global expression of lacZ in the CNS (data not shown).

Especially for larval stages, postembryonic NBs and their progeny cells have been labelled by exposure to [3H]thymidine (White and Kankel, 1978) or BrdU (Truman and Bate, 1988). By these methods, the large cells appearing peripherally in the CNS cortex of the larva (10-12qm, White and Kankel, 1978; 8 – 10 μm, Truman and Bate, 1988 ; 7 – 11 μm, own observations) were identified as postembryonic NBs and the nests of cells lying beneath to be their progeny. This situation appears to be conserved among several holometabolous insects (Bauer, 1904; Nordlander and Edwards, 1969; Booker and Truman, 1987).

Histochemical staining of the clones resulting from our transplantations of early embryonic precursor cells was carried out in the late third instar larva. The purely blue subclones consisting of a large peripheral cell and an associated nest of small cells correspond in morphology and position to the descriptions mentioned above. Whereas all embryonic descendants of the transplants are strongly labelled with HRP (Technau, 1986), increase in volume (about 6-fold) of postembryonic NBs in the larva and subsequent proliferation leads to an almost complete dilution of HRP in these cells and their progeny (Tix et al. 1989a,b). Therefore, in our preparations, a purely blue-staining subclone represents a postembryonic NB and its progeny cells.

Coinjection of the fluorescent lineage tracer RITC-dextran allowed us to inspect the clones in vivo at late embryonic stages, i.e., before the postembryonic NBs appear. It was not possible to resolve the fluorescently labelled clones in every detail but nevertheless their position and approximate cell numbers were identifiable and in all cases corresponded to the HRP-labelled parts of the same clones subsequently stained in the late larva. This, and the fact that the HRP-labelled cells are fully differentiated, whilst differentiation of postembryonically derived cells is still at an early stage, indicates that they represent components of the larval CNS that developed during embryogenesis.

Occasionally, a few cells at the interface between the embryonic and postembryonic part of the clone may stain pale brown (Fig. 3A). Since they clearly carry a much lower concentration of the label than the other HRP-labelled cells, we interpret these cells as being early postembryonic ones. In these cases, the tracer may not have become sufficiently diluted out during enlargement of the postembryonic NB, so that residual HRP is transferred to its first progeny cells before it dilutes out completely during subsequent divisions.

In the thoracic neuromeres, the number of postem-bryonic NBs has been shown to be similar to the number of embryonic NBs (Truman and Bate, 1988). This does not conform with the results from our transplantation experiments. Only 68 % of the resulting clones contained embryonic as well as postembryonic progeny cells, whilst 32 % consisted of embryonic cells only. These purely embryonic clones were not confined to particular regions within the CNS. For example, one might expect to find such clones near the midline as there are progenitor cells there that behave unlike NBs (Bate and Grünewald, 1981; Thomas et al. 1984; Crews et al. 1988). Instead, these clones also appeared in all other ventrodorsal positions at similar frequencies. Furthermore, in all clones with postembryonic cells, we could only detect one NB. Thus, the high rate of clones that did not have postembryonic subclones is not compensated for by other clones carrying more than one postembryonic NB.

The fact that many clones did not contain postembryonically derived cells could be attributed to a variety of causes. First of all, artefacts of the transplantation procedure might prevent some of the NBs from resuming their mitotic activity during postembryonic development and cause them to behave like most of the abdominal NBs. Secondly, since we removed up to 10 cells at a time from the donors, we cannot exclude that occasionally a few cells from the presumptive abdominal region of the neurogenic ectoderm may have entered the capillary. In the embryo, the number of embryonic NBs is nearly the same in all segments (Hartenstein and Campos-Ortega, 1984). In the larva, however, abdominal neuromeres contain about eight times less NBs than the thoracic neuromeres (Truman and Bate, 1988). If the precursor cells were already firmly committed as thoracic or abdominal cells at the early gastrula stage (Chan and Gehring, 1971; Simcox and Sang, 1983), they would develop according to their origin upon heterotopic transplantation along the anterior-posterior axis. Determination of segmental specificity of precursor cells will be the subject of further experiments.

Identity of embryonic and postembryonic neuroblasts

We found two types of clones in thoracic ganglia. One consisted of embryonically derived cells only, the other type contained a postembryonic subclone in addition. We never found clones consisting exclusively of postembryonically derived cells.

The fact that postembryonic progeny always appear in one clone with embryonic progeny indicates that the postembryonic NBs are derived from the same precursor cells as embryonic NBs. But this does not necessarily imply that both sets of NBs represent identical cells (see B2 in Fig. 1). It could also be that one of the progeny cells of the embryonic NB is set aside to become the postembryonic NB whereas the embryonic NB degenerates or is used up by its last division in the embryo (see Bl in Fig. 1). Based on the observation that numbers of NBs in the thoracic neuromeres are similar in the embryo and in the larva, Truman and Bate (1988) discussed the possibility that embryonic and postembryonic NBs may be the same cells. If this were right, it should be possible to label postembryonic NBs with BrdU during their last division as embryonic NBs. According to Hartenstein et al. (1987) embryonic NBs perform 8-9 parasynchronous mitotic cycles between stages 9 and 14. Our data suggest that proliferation of a subpopulation of embryonic NBs lasts until late stage 16. In the vNC of larvae that were injected with BrdU at late stage 16, a subpopulation of postembryonic NBs in the thorax stains positive for BrdU. They are generally associated with two or sometimes three progeny cells which had been labelled during embryogenesis. Since with each division NBs give rise to a smaller ganglion mother cell which in turn divides once to produce two larval neurones (Bauer, 1904; Bate, 1976; Hartenstein et al. 1987), it is likely that injections at late stage 16 preceded the last mitosis of embryonic NBs. Thus, this experiment lends support to the assumption that embryonic and postembryonic NBs are identical cells (see B2 in Fig. 1), although it does not exclude the possibility that postembryonic NBs emerge from one of the latest embryonic ganglion mother cells.

Since the majority of embryonic NBs stop dividing at earlier embryonic stages, injections at earlier stages lead to an increased number of labelled postembryonic NBs as well as progeny cells. Thus, depending on when the injections are performed, different sets of NBs become labelled during their last embryonic DNA synthesis. But, for several reasons the patterns derived from BrdU injections at these earlier stages are difficult to interpret. Therefore, we are unable to decide whether the above interpretation may be true for all imaginai NBs.

Our data present evidence that postembryonic NBs in Drosophila have the same origin as embryonic NBs. They further suggest that at least some of the embryonic and postembryonic NBs are identical. Therefore, holometabolous insects seem to have acquired a regulative mechanism that allows mitosis of NBs to be interrupted at the interface between embryonic and postembryonic development.

We would like to thank Michael Bate for valuable comments on the manuscript, Jose A. Campos-Ortega for discussions, and Yuh-Nung Jan, Cahir O’Kane and Ross J. MacIntyre for providing fly strains. This work was supported by a grant from the Deutsche Forschungsgemeinschaft to G.M.T.

Bate
,
C. M.
(
1976
).
Embryogenesis of an insect nervous system. I. A map of the thoracic and abdominal neuroblasts in Locusta migratoria
.
J. Embryol. exp. Morph
.
35
,
107
123
.
Bate
,
C. M.
and GR
Fc;newald
,
E. B.
(
1981
).
Embryogenesis of an insect nervous system. II. A second class of neuron precursor cells and the origin of the intersegmental connectives
.
J. Embryol. exp. Morph
.
61
,
317
330
.
Bauer
,
V.
(
1904
).
Zur inneren Metamorphose des Zentralnervensystems der Insekten
.
Zool. Jhb. Anat. Ontog. Tiere
20
,
123
150
.
Bellen
,
H. J.
,
O’kane
,
C. J.
,
Wilson
,
C.
,
Grossniklaus
,
U.
,
Pearson
,
R. K.
and
Gehring
,
W. J.
(
1989
).
P-element-mediated enhancer detection: a versatile method to study development in Drosophila
.
Genes and Dev
.
3
,
1288
1300
.
Bier
,
E.
,
Vâssin
,
H.
,
Shepherd
,
S.
,
Lee
,
K.
,
Mccall
,
K.
,
Barbel
,
S.
,
Ackerman
,
L.
,
Carreteo
,
R.
,
Uemura
,
T.
,
Grell
,
E.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1989
).
Searching for pattern and mutation in the Drosophila genome with a P-lac Z vector
.
Genes and Dev
.
3
,
1273
1287
.
Booker
,
R.
and
Truman
,
J. W.
(
1987
).
Postembryonic neurogenesis in the CNS of the Tobacco Homworm, Manduca sexta. I. Neuroblast arrays and the fate of their progeny during metamorphosis
.
J. comp. Neur
.
255
,
548
559
.
Campos-Ortega
,
J. A.
and
Hartenstein
,
V.
(
1985
).
The Embryonic Development of Drosophila melanogaster
.
Springer
,
Berlin Heidelberg New York Tokyo
.
Chan
,
L. N.
and
Gehring
,
W. J.
(
1971
).
Determination of blastoderm cells in Drosophila melanogaster
.
Proc. natn. Acad. Sci. U.S.A
.
68
(
9
),
2217
2221
.
Crews
,
S. T.
,
Thomas
,
J. B.
and
Goodman
,
C. S.
(
1988
).
The Drosophila single minded gene encodes a nuclear protein with sequence similarity to the per gene product
.
Cell
52
,
143
151
.
Doe
,
C. Q.
and
Goodman
,
C. S.
(
1985
).
Early events in insect neurogenesis. I. Development and segmental differences in the pattern of neuronal precursor cells
.
Devi Biol
.
Ill
,
193
205
.
Gratzner
,
H. G.
(
1982
).
Monoclonal antibody to 5-bromo- and 5-iododeoxyuridine: a new reagent for detection of DNA replication
.
Science
218
,
474
f.
Hartenstein
,
V.
and
Campos-Ortega
,
J. A.
(
1984
).
Early neurogenesis in wild-type Drosophila melanogaster
.
Roux’ Arch, devl Biol
.
193
,
308
325
.
Hartenstein
,
V.
,
Rudloff
,
E.
and
Campos-Ortega
,
J. A.
(
1987
).
The pattern of proliferation of the neuroblasts in the wild-type embryo of Drosophila melanogaster
.
Roux’s Arch, devl Biol
.
196
,
473
485
.
Hartenstein
,
V.
,
Technau
,
G. M.
and
Campos-Ortega
,
J. A.
(
1985
).
Fate-mapping in wild-type Drosophila melanogaster. III. A fate map of the blastoderm
.
Roux’s Arch, devl Biol
.
194
,
213
216
.
Hertweck
,
H.
(
1931
).
Anatomie und Vanabilitât des Nervensystems und der Sinnesorgane von Drosophila melanogaster (Meigen)
.
Z. wiss. Zool
.
139
,
559
663
.
Kankel
,
D. R.
and
Hall
,
J. C.
(
1976
).
Fate mapping of nervous system and other internal tissues in genetic mosaics of Drosophila melanogaster
.
Devl Biol
.
48
,
1
24
.
Knipple
,
D. C.
and
Macintyre
,
R. J.
(
1984
).
Cytogénie mapping and isolation of mutations of the /LGal-1 locus of Drosophila melanogaster
.
Molec. gen. Genet
.
198
,
75
83
.
Malzacher
,
P.
(
1968
).
Die Embryogenèse des Gehims paurometaboler Insekten. Untersuchungen an Carausius morosas und Periplaneta americana
.
Z. Morph. Okol. Tiere
62
,
103
161
.
Nordlander
,
R. H.
and
Edwards
,
J. S.
(
1969
).
Postembryonic brain development in the Monarch Butterfly Danaus plexippus plexippus, L. I. Cellular events during brain morphogenesis
.
W’dhelm Roux’s Arch. Entwmech. Org
.
162
,
197
217
.
O’kane
,
C.
and
Gehring
,
W. J.
(
1987
).
Detection in situ of genomic regulatory elements in Drosophila. Proc. natn
.
Acad. Sci. U.S.A
.
84
,
9123
9127
.
Poulson
,
D. F.
(
1950
).
Histogenesis, organogenesis and differentiation in the embryo of Drosophila melanogaster, Meigen
. In
Biology of Drosophila
(ed.
M.
Demerec
).
Wiley
,
New York
, pp.
168
274
.
Shepherd
,
D.
and
Bate
,
C. M.
(
1990
).
Spatial and temporal patterns of neurogenesis in the embryo of the locust (Schistocerca gregarina)
.
Development
108
,
83
96
.
Simcox
,
A. A.
and
Sang
,
J. H.
(
1983
).
When does determination occur in Drosophila embryos?
Devl Biol
.
97
,
212
221
.
Technau
,
G. M.
(
1986
).
Lineage analysis of transplanted individual cells in embryos of Drosophila melanogaster. I. The method
.
Roux’s Arch, devl Biol
.
195
,
389
398
.
Technau
,
G. M.
(
1987
).
A single cell approach to problems of cell lineage and commitment during embryogenesis of Drosophila melanogaster
.
Development
100
,
1
12
.
Technau
,
G. M.
and
Campos-Ortega
,
J. A.
(
1986
).
II. Commitment and proliferative capabilities of neural and epidermal cell progenitors
.
Roux’s Arch, devl Biol
.
195
,
445
454
.
Thomas
,
J. B.
,
Bastiani
,
M. J.
,
Bate
,
C. M.
and
Goodman
,
C. S.
(
1984
).
From Grasshopper to Drosophila-, a common plan for neuronal development
.
Nature
310
(
19
),
203
207
.
Tix
,
S.
,
Bate
,
C. M.
and
Technau
,
G. M.
(
1989a
).
Pre-existing neuronal pathways in the developing leg imaginai discs of Drosophila
.
Development
107
,
855
862
.
Tix
,
S.
,
Minden
,
J. S.
AND
Technau
,
G. M.
(
1989b
).
Pre-existing neuronal pathways in the developing optic lobes of Drosophila
.
Development
105
,
739
746
.
Truman
,
J. W.
and
Bate
,
C. M.
(
1988
).
Spatial and temporal patterns of neurogenesis in the CNS of Drosophila melanogaster
.
Devl Biol
.
125
,
145
157
.
Weismann
,
A.
(
1864
).
Die nachembryonale Entwicklung der Musciden nach Beobachtungen an Musca vomitoria und Sarcophaga carnaria
.
Z. wiss Zool
.
14
,
187
336
.
Wheeler
,
W. M.
(
1891
).
Neuroblasts in the arthropod embryo
.
J. Morph
.
4
,
337
343
.
White
,
K.
and
Kankel
,
D. R.
(
1978
).
Patterns of cell division and cell movement in the formation of the imaginai nervous system in Drosophila melanogaster
.
Devl Biol
.
65
,
296
321
.