Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Mechanisms of Development 74 (1998) 99–110 Homeotic regulation of segment-specific differences in neuroblast numbers and proliferation in the Drosophila central nervous system Andreas Prokop a ,*, Sarah Bray b, Emma Harrison b, Gerhard M. Technau a a Institut für Genetik-Zellbiologie, Becherweg 32, D-55128 Mainz, Germany b Department of Anatomy, Downing Street, Cambridge CB2 3DY, UK Received 9 February 1998; revised version received 20 April 1998; accepted 20 April 1998 Abstract The number and pattern of neuroblasts that initially segregate from the neuroectoderm in the early Drosophila embryo is identical in thoracic and abdominal segments. However, during late embryogenesis differences in the numbers of neuroblasts and in the extent of neuroblast proliferation arise between these regions. We show that the homeotic genes Ultrabithorax and abdominal-A regulate these late differences, and that misexpression of either gene in thoracic neuroblasts after segregation is sufficient to induce abdominal behaviour. However, in wild type embryos we only detect abdominal-A and Ultrabithorax proteins in early neuroblasts. Furthermore, transplantation experiments reveal that segment-specific behaviour is determined prior to neuroblast segregation. Thus, the segment-specific differences in neuroblast behaviour seem to be determined in the early embryo, mediated through the expression of homeotic genes in early neuroblasts, and executed in later programmes controlling neuroblast numbers and proliferation. 1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: BrdU; Central nervous system; Drosophila; Homeotic genes; Proliferation; Transplantation 1. Introduction The central nervous system (CNS) is composed of a huge variety of cells, most of which are unique in their properties. In insects these individual neurons are arranged in stereotyped patterns with reproducible differences between the segments corresponding to the diverse regional requirements of the CNS, such as the control of the locomotion apparatus in the thorax or the reproduction organs in the abdomen. It is thus important to understand how regional differences are programmed during nervous system development. The ventral nerve cord in insects derives from neural stem cells, the neuroblasts (NBs), which segregate in reproducible segmental patterns from the ventral neurogenic region of the ectoderm (Bate, 1976; Hartenstein and Campos-Ortega, 1984; Doe, 1992). In Drosophila, individual NBs give rise to lineages of specific size and cell composition (Udolph et al., 1993; Bossing and Technau, 1994; * Corresponding author. Tel.: +49 6131 394328; fax: +49 6131 395845; e-mail: [email protected] 0925-4773/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0925-4773 (98 )0 0068-9 Bossing et al., 1996; Schmidt et al., 1997). However, although the pattern of segregating NBs is identical in thoracic and abdominal segments (Doe, 1992), reproducible differences occur between the thoracic and abdominal versions of some NB lineages (Prokop and Technau, 1994a; Bossing et al., 1996; Schmidt et al., 1997). This is also reflected in the proliferation patterns in the CNSs which show marked differences between thorax and abdomen at stages 15 and 16 (Prokop and Technau, 1991). Segment-specific differences are also evident in the number of NBs that persist beyond the end of embryogenesis and proliferate during larval stages. At stage 17, all NBs have stopped dividing but can still be monitored by NB-specific expression of grainyhead (grh) (Bray et al., 1989). Analyses of GRH expression patterns in the CNSs of wild type embryos and of mutant embryos where cell death is suppressed strongly suggest that a number of NBs normally die towards the end of embryogenesis (White et al., 1994). The degree of cell death shows segment-specific differences, in that many more NBs die in the central abdomen than in the thorax and anterior abdomen. As a consequence, when NBs 100 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 resume proliferation as postembryonic NBs in the larval stages, 47 NBs are detected in each thoracic, about 12 in the two anterior abdominal neuromeres, but only six in central abdominal segments. Furthermore, postembryonic NBs in the thorax and anterior abdomen produce hundreds of daughter cells each, whereas those in abdominal neuromeres A3–A7 give rise to only five to 15 cells (Truman and Bate, 1988; Prokop and Technau, 1994b). In summary, there are three major factors regulating the segment-specific proliferation of NBs: (1) the period and frequency of embryonic NB proliferation, (2) the number of NBs eliminated at the end of embryogenesis, and (3) the frequency and period of postembryonic proliferation. The development of segment-specific characteristics is programmed during development by the homeotic genes which encode homeodomain transcription factors and are conserved from nematodes to vertebrates, both with respect to their function and to their organisation within gene complexes (for reviews see Lewis, 1978; Duncan, 1987; Beachy, 1990; Kaufman et al., 1990; McGinnis and Krumlauf, 1992; Morata, 1993). Homeotic selector genes are in principle expressed in those body regions in which they are required to select the developmental pathways specific for each particular segment. In Drosophila, mutations in any of the homeotic genes result in the transformation of the morphological characteristics of the segments, where they are normally expressed, into those of other (in general more anterior) segments. Following these principles, segment identities in the head and anterior thorax of Drosophila are controlled by genes of the Antennapedia-complex, whereas the posterior thorax and the abdomen are controlled by genes of the bithorax-complex, including Ultrabithorax (Ubx) and abdominal-A (abd-A). The homeotic selector genes are expressed in a complex pattern in the Drosophila CNS (Doe et al., 1988) and the segment-specific development of at least one embryonic NB lineage is regulated by homeotic genes (Prokop and Technau, 1994a). The bithorax-complex genes are thus likely candidates for regulating the thoracic and abdominal patterns of NB behaviour. To investigate this possibility we have analysed whether loss-of function mutations or misexpression of homeotic genes have any effects on (a) patterns of bromo deoxyuridine (BrdU) incorporation which indicates DNA synthesis and is used to give an indication of NB proliferation (Truman and Bate, 1988; Prokop and Technau, 1991), and (b) late embryonic patterns of NBspecific GRH expression, which labels the persisting NBs at that stage (White et al., 1994; see above). Our analysis indicates that abd-A and Ubx regulate the differences in cell number between thoracic and abdominal neuromeres that are seen in the larval and adult CNS by altering both the amount of proliferation in embryo and larva, and the number of NBs that are present in post-embryonic stages. Although segment-specific characteristics of BrdU and GRH patterns are first observed towards the end of embryogenesis, we only detect the expression of ABD-A and UBX in NBs during earlier embryonic stages. Furthermore, transplantation experiments indicate that determination of thoracic versus abdominal characteristics occurs prior to NB segregation, before the homeotic proteins themselves are expressed. We propose therefore that segment-specific differences in neuroblast behaviour are determined during the early patterning of the embryo which results in specific expression of homeotic genes in early NBs. This confers later programmes of proliferation control, which no longer require the activity of homeotic genes. 2. Results 2.1. Segment-specific differences in late embryonic NB proliferation are controlled by abd-A and Antennapedia Individual NBs in the Drosophila CNS undergo different patterns of division depending on their position (Prokop and Technau, 1994a; Bossing et al., 1996; Schmidt et al., 1997). Patterns of replicating cells can be detected through their incorporation of BrdU which in late stages of embryogenesis reveals scattered cells in the abdominal neuromeres compared with regular incorporation patterns in the more anterior regions of the ventral nerve cord (Fig. 1B,C). In the subesophageal ganglion approximately 20 groups of two to four cells in ventral to ventrolateral locations are labelled, and on either side of each thoracic neuromere one to three groups of two to four cells are detected in ventrolateral/ lateral (but not in ventral) positions. Given that the initial number and organisation of NBs is the same in thorax and abdomen (Doe, 1992), the observed replication patterns indicate that equivalent NBs behave differently in the thorax and the abdomen; in the thorax they continue replicating whereas in the abdomen they do not. In order to determine whether homeotic genes regulate these differences between thoracic and abdominal DNA replication, BrdU was injected into homeotic mutant embryos at stage 16/17. In mutant embryos lacking Ubx and abd-A function (Df109), BrdU incorporation appears normal in the subesophageal ganglion and in the thorax. However, in the abdominal neuromeres (except for the terminal region) lateral cells incorporate BrdU in patterns reminiscent of the thoracic region (not shown, but see Fig. 1E). Mutant embryos lacking Ubx function alone, show no obvious changes in the pattern of BrdU incorporation (not shown) in spite of a clear presence of the homeotic cuticle phenotype (Lewis, 1978). In contrast, embryos lacking abdA function had defects reminiscent of embryos carrying the deficiency Df109 (Fig. 1E) with extra lateral NBs incorporating BrdU in abdominal segments. The regulation by abd-A but not Ubx suggests that the NBs lie in the posterior compartment of the segment, because the effects are also seen in A1 where abd-A is only expressed in the posterior compartment (see Fig. 6A). The posterior localisation of the lateral thoracic NB in wild type embryos was confirmed by double A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 labelling with antibodies against BrdU and ENGRAILED, a marker for the posterior compartment (Fig. 1D). In embryos lacking the Antennapedia gene, thoracic BrdU incorporation associated with ventrolateral/lateral NBs is normal, however, additional staining is detected in ventral positions resembling the ventral BrdU patterns of the subesophageal ganglion (Fig. 1F). Taken together, these results demonstrate that in late embryos abd-A function is needed to repress DNA replication in some lateral NBs of abdominal neuromeres, and Antennapedia function is required to repress DNA replication in ventral NBs of the thorax. As pulse chase experiments have shown that the pattern of BrdU incorporation in the thoracic neuromeres of late 101 embryos reflects NB proliferation (Prokop and Technau, 1991) we believe that the DNA replication patterns we observe in these experiments are indicative of effects on NB proliferation. 2.2. Homeotic genes also control the numbers and proliferation of postembryonic NBs During the second larval instar, some of the embryonic NBs increase in size and resume proliferation as postembryonic NBs (pNBs). The thoracic neuromeres contain 47 pNBs and the abdominal neuromeres A1 and A2 about 12 pNBs, all of which proliferate extensively. In contrast, the Fig. 1. Homeotic regulation of segment-specific BrdU-incorporation in the embryonic CNS. Late stage 17 embryos in lateral (A,B,J) or ventral view (C–H; anterior always to the left), labelled with antibodies against ENGRAILED (EN) and/or BrdU as indicated (bottom right). (A) Each engrailed stripe indicates one neuromere (T, thorax, delimited by black lines; Ab, abdomen; H, hemispheres; S, subesophageal ganglion). (B–D) White arrowheads point at lateral BrdU incorporation in the thorax, black arrows at ventral proliferation in the subesophageal ganglion, black arrowheads at engrailed positive cells. Note that BrdU labelled cells lie in one line with engrailed positive cells, thus in the posterior neuromere compartment. (E–G) show mutant embryos as indicated (top middle), (H,J) are embryos with misexpression of UBX; white arrows indicate lateral BrdU incorporation comparable with wild type (see (B,C)), open arrowheads indicate ectopic lateral, open arrows ectopic ventral and a bent arrow loss of lateral BrdU incorporation. (K) Ectopic UBX expression appears in the NB layer about stage 10/11 (white arrow), but not in the peripheral ectoderm (except intrinsic expression in T2 and T3; asterisks). Subesophageal (S) and thoracic (T1–3) neuromeres are indicated. Scale bar, 50 mm (A–C, E–J); 20 mm (D,K). 102 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 abdominal neuromeres A3–7 contain at most six pNBs (two sets of vm-, vl- and dl-pNBs), each of which divides only a few times (Fig. 2A,B; Truman and Bate, 1988; Prokop and Technau, 1994b, 1995). In order to investigate a possible involvement of homeotic genes in the regulation of these postembryonic segment-specific differences we analysed patterns of BrdU incorporation or of pNB-specific toluidine blue staining (see Truman and Bate, 1988) in allelic combinations of homeotic genes that allow mutant animals to survive into larval stages (Lewis, 1978; Ghysen et al., 1985). Larvae carrying just one copy of the Ubx and abdA genes (Df109/+), show additional large clusters of BrdU labelled cells in A3, A4 and sometimes also A5 (not shown, but see Fig. 2G). Most of these clusters are located in a ventromedial position derived either from the vm-pNBs or from additional adjacent pNBs. Sometimes further BrdU labelled large cell clusters are located in more dorsal positions, adjacent to or originating from the vl- or dl-pNB. Larvae carrying just one copy of the Ubx gene have slight defects in the BrdU pattern in one-fourth of cases, with one larger cell group in ventrolateral positions of A3 and/or an enlargement of the vm-cluster (normally about five cells) to up to 15 cells in A3 or A4 (not shown). However, these defects are less severe than in larvae heterozygous for Df109. In contrast, larvae lacking one copy of the abd-A gene, all show a BrdU pattern resembling the severe defects observed in larvae carrying the Df109 deletion (Fig. 2G). This demonstrates that abd-A is the principle factor required to regulate segment-specific features of pNB proliferation in most of the abdominal neuromeres, comparable with the situation in the embryonic CNS. This interpretation is supported by analyses of larvae carrying one copy of the abd-A allele Uab4 (Lewis, 1978) in combination with an abd-A null allele (Uab4/Df109 or Uab4/abd-AMX1). Such larvae have about 12 pNBs in each of the abdominal neuromeres A3–A7 (compared with six pNBs in the wild type; Fig. 2A,D). Each of these surplus pNBs produces a hundred or more progeny, whereas in wild type larvae the pNBs in A3–A7 produce at most 15 progeny (Fig. 2B,E; Truman and Bate, 1988; Prokop and Technau, 1994b). Both the number and the proliferation behaviour of these pNBs in the central abdomen of the mutant larvae is reminiscent of the wild type pattern described for neuromere A2 (Truman and Bate, 1988). The NB pattern defects can already be detected in the late embryo (stage 16/17) using the NB-specific marker grainyhead (grh): many more GRH-expressing pNBs persist in the abdominal neuromeres of Uab4/Df109 mutant embryos than in wild type (Fig. 2C,F). Thus, abd-A regulates the proliferation of pNBs both by controlling the number of NBs that are eliminated in the late embryo and by regulating the number of divisions the persisting NBs undergo during larval stages. Fig. 2. Effects of homeotic mutations on the pattern and proliferation of postembryonic NBs. (A,B,D,E,G) BrdU incorporation in nerve cords of late third instar larvae exposed to BrdU during larval life. (C,F,H,I) Anti-GRH labelled nerve cords at late stage 17. Anterior is to the left. (A) In the wild type larva, abdominal neuromeres A3–A7 (delimited by lines) have six clusters of BrdU labelled cells derived from two pairs of vm-, vl- and dl-pNBs, respectively (see higher magnification in (B)), comparable with the abdominal NB patterns in late embryos (C); arrows in (A–C) indicate vm-pNBs. In Uab4/abd-A or Uab4/ DF109 mutant larvae many more NBs persist in A3–A7 (D,F) and produce larger lineages (E). (G) In abd-A/+ mutant larvae A3, A4 and sometimes A5 have few large NB lineages (black arrowheads). Misexpression of UBX (H) or ABD-A (I) in all NBs removes most thoracic NBs (white arrowheads indicate remaining NBs; arrows indicate hemispheres). Scale bar, 20 mm (F); 30 mm (B,E); 35 mm (C,H,I); 100 mm (A,D,G). A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 2.3. Misexpression of bithorax-complex genes is sufficient to confer segment-specific behaviour on NBs Impaired function or reduced levels of bithorax-complex genes affect NB behaviour in the domain where these genes are normally expressed, consistent with their direct requirement for the regulation of abdominal NB characteristics. In order to test whether the presence of ABD-A or UBX is sufficient to confer abdominal NB behaviour we analysed the consequences of expressing these proteins in all NBs including those in thoracic neuromeres using the GAL4expression system (Brand and Perrimon, 1993). When the transcription of UAS-Ubx or UAS-abd-A was ectopically induced in most or all segregated embryonic NBs and their progeny (but not in the neuroectoderm; Fig. 1K) with the GAL4 driver line MZ1407 the overall shape of the CNS was not perturbed indicating that most phases of proliferation are not grossly abnormal. However, the pattern of BrdU incorporation at stage16/17 is altered with misexpression of either UBX or ABD-A abolishing the lateral replicating clusters in the thoracic neuromeres (Fig. 1J; bent arrow). This is what normally occurs in the abdominal NBs suggesting that the thoracic expression of ABD-A or UBX after NB segregation is sufficient to transform the characteristics of some thoracic lateral NBs into those of abdominal NBs. However, we also find ectopic sites of BrdU-incorporation in the ventral region of thoracic neuromeres (Fig. 1H,J). The pattern appears like an extension of the subesophageal pattern and resembles the phenotype of Antennapedia mutant embryos (Fig. 1F vs. H,J; open arrows), as if ectopic bithorax gene activity in the thoracic neuromeres represses Antennapedia function. Indeed ectopic UBX has been shown to repress Antennapedia gene expression (González-Reyes and Morata, 1990). Thus, our results indicate that specific NBs respond differently to ectopic expression of bithorax-complex genes, acquiring abdominal characteristics in some cases and subesophageal characteristics in others. Similar results are obtained in Polycomb mutant embryos, which have ectopic expression of ABD-A and UBX in the NB layer from stage 10/11 onwards (Simon et al., 1992; Prokop and Technau, 1994a). In Polycomb3 mutant embryos at stage 16, both the lack of lateral and gain of ventral BrdU incorporating cells are observed, similar to the GAL4 experiments (Fig. 1G). Therefore, our data are consistent with the hypothesis that expression of bithorax-complex genes after NB segregation is sufficient to regulate the segment-specific replication of embryonic NBs. The homeotic genes do not directly interfere with the cell cycle (e.g. block it completely upon misexpression), but they appear to determine which segmental proliferation schedule is carried out. As only a few NB lineages have segment-specific BrdUincorporation patterns in the embryo we wanted to investigate whether misexpression of ABD-A or UBX is also sufficient to alter the numbers of NBs at the end of embryogenesis, which is a characteristic relevant to many 103 more NB lineages. Misexpression of either Ubx or abd-A leads to a reduction in the number of grh-expressing cells in the CNSs of stage 17 embryos which we take to indicate that they affect the number of NBs that are eliminated (see Section 1; White et al., 1994). The effects are dramatic in the thoracic neuromeres where only four to eight grh-expressing cells remain per neuromere at specific locations (Fig. 2H,I). Thus, the effects of misexpressing ABD-A and UBX on the patterns of BrdU-incorporation and GRH expression in the late embryo suggest that most, if not all, thoracic NBs are responsive to bithorax-complex genes, and that the presence of ABD-A or UBX after NB segregation is sufficient to induce abdominal behaviour. 2.4. Expression of homeotic genes in NBs precedes the appearance of segment-specific features To determine when bithorax gene function is required during embryogenesis to regulate NB behaviour, we investigated when these proteins are present in the abdominal NBs and therefore likely to impose their regulation. Using anti-GRH or a grh-lacZ reporter-gene (which recapitulates the NB expression of grainyhead) to mark the NBs we detected no co-expression with ABD-A in the persisting NBs of stage 17 embryos (not shown). In stage 15 embryos prior to NB elimination, many more cells express grh. However, even at this stage there are no cells which co-express UBX and GRH and only rarely (one to two ventral cells per neuromere) are ABD-A and GRH expression detected in the same cells (Fig. 3D,E). This suggests that neither ABD-A nor UBX are present in the NBs shortly before segmentspecific differences in BrdU and GRH patterns occur. During early stages of embryogenesis NB can be detected either by their position and shape or by expression of asense. ASENSE is expressed in NBs shortly after their segregation from the neuroectoderm and so can be used as an early marker for NBs. UBX was detected in many NBs at stages 8–12 although there is wide variation between the levels of UBX present in different NBs and a subset of NBs contain no detectable UBX (Fig. 3A,C). Similarly, ABD-A is present in many NBs at early stages (Fig. 3B). Thus, both UBX and ABD-A are expressed in embryonic NBs, but their expression fades before segment-specific differences become detectable. To see whether we could define a precise stage when homeotic gene function is sufficient to confer segment-specific NB elimination and proliferation patterns we tested the effects of inducing ABD-A or UBX at different stages via heat inducible promoters. None of the heat shock regimes were able to confer homeotic changes on proliferation or GRH patterns (although these experiments clearly caused homeotic effects in the epidermis), so we cannot identify the precise stage when the activity of these genes is required. Nevertheless the timing of ABD-A and UBX expression in NBs makes it unlikely that the homeotic genes directly regulate those genes which mediate proliferation cessation or NB elimination. Instead they seem to con- 104 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 Fig. 3. ABD-A and UBX are found in NBs only at earlier embryonic stages. ABD-A or UBX expression (as indicated top right) in embryos at stage 10/11 (A–C) or stage 15 (D,E), or (F) in early third instar larval CNSs (L3; stages indicated bottom right). NBs were identified by position and shape (A,B,F) or expression of ASENSE (ASE; (C)) or GRH (D,E). (A–C) NBs contain homeotic proteins at early stages (white arrowheads; orange colour in (C) indicates double labelling). (D–F) NBs lack homeotic gene expression later on (white arrowheads; for L3 see also Truman et al., 1993). Scale bar, 15 mm (A,B,F); 35 mm (C–E). trol regulatory pathways, which function subsequent to the expression of the bithorax-complex genes. 2.5. Thoracic versus abdominal behaviour of NB is determined prior to NB segregation The misexpression data indicate that homeotic genes determine segment-specific features in the CNS after NBs have segregated. However, the analysis of the NB1-1 lineage revealed that segment-specificity can be determined in the peripheral ectoderm, upstream of homeotic gene function (Prokop and Technau, 1994a). We therefore tested whether the segment-specific postembryonic proliferation patterns are also determined in the peripheral ectoderm, to see whether this early determination could be a general mechanism underlying CNS development. Single cells, labelled genetically with the gene encoding b-galactosidase and cytoplasmically with HRP (horseradish peroxidase), were transplanted between the thoracic and the abdominal neuroectoderm at the early gastrula stage about 30 min prior to the onset of NB segregation. As b-galactosidase is genetically expressed by every cell of the clone, both the embryonically and postembryonically derived cells will contain bgalactosidase activity. However, only the embryonically derived cells will contain HRP, as the enzyme gets diluted below detectable level when the NB takes up proliferation in the larva (Fig. 4B,D; for details see Prokop and Technau, 1991). The composition of the clones was analysed in the late larval ventral nerve cord when the differences between abdominal and thoracic NB lineages are very distinctive. We obtained 66 CNS clones from isotopic transplantations in the thorax, 45 of which contained a pNB with a nest of daughter cells exclusively labelled by b-galactosidase activity indicating that they were postembryonic progeny (Figs. 4A,B and 5, T to T). This proportion of 68% is in accordance with the number of thoracic NBs which are reactivated during larval life in relation to the total number of embryonic precursors (47 pNBs versus 60 NBs and seven to eight midline precursors per neuromere in the embryo; Truman and Bate, 1988; Doe, 1992; Bossing and Technau, 1994). Isotopic transplantations in the abdomen yielded 32 abdominal CNS clones of which only two contained a larger peripheral cell (potential pNB) and a small cluster of postembryonic progeny cells. This low rate of NB reactivation (6%) and their low proliferation rate reflect the typical behaviour of abdominal pNBs (Fig. 5, A to A). When cells were heterotopically transplanted from thoracic to abdominal sites of the early gastrula neuroectoderm, 67% gave rise to a large nest of postembryonic cells with a pNB (Figs. 4C,D and 5, T to A) consistent with the characteristics of thoracic NBs. Conversely, when cells were transplanted from abdominal to thoracic sites, all clones failed to express thoracic features and contained only (embryonic) cells strongly labelled by HRP activity (Fig. A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 105 Fig. 4. Examples of late larval NB lineages originating from transplantations at the early gastrula stage. (A–D) Thoracically derived pNBs (arrowheads) with embryonic (brown cells) and postembryonic (blue cells) progeny can develop in the thorax (A,B) or abdomen (C,D). (E) Lineages without pNBs always lack large nests of purely blue cells. (F) Schematic representation of an embryo at the early gastrula stage indicating the modes of transplantation that led to the lineages in (A–E). Scale bar, 70 mm (A,C); 20 mm (B,D). 4E and 5, A to T). Thus, NBs autonomously execute postembryonic proliferation patterns according to their domain of origin in the neuroectoderm. This indicates that most or even all neuroectodermal CNS precursors are determined at the early gastrula stage with respect to their thoracic or abdominal identity. Since subsequent misexpression of homeotic genes can alter the behaviour of NBs, our findings suggest that genetic mechanisms upstream of homeotic gene function must initially determine the regional identity of these cells. 3. Discussion 3.1. Homeotic genes regulate proliferation and numbers of NBs The development of the nervous system involves the gen- eration of the correct numbers of neural precursors, their subsequent division to generate a defined set of progeny and the acquisition by their progeny of individual neural cell fates. These processes must be differentially modulated along the body axis so that the neural structures appropriate for each region of the body develop. Previous studies in Drosophila have emphasised the role of homeotic genes in controlling the types of neural fate that are adopted by the progeny of the NBs (Doe and Scott, 1988). Our results demonstrate that the homeotic genes also regulate segment-specific differences in NB numbers and NB divisions. This regulation is achieved both by controlling the proliferation period and frequency and by regulating whether NBs persist at the end of embryogenesis. Because of the latter, the number and pattern of postembryonic NBs depend on homeotic gene function unlike the embryonic NBs where the pattern of segregating NBs is the same between thoracic and abdominal neuromeres (Doe, 1992). Thus in the larval 106 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 Fig. 5. Isotopic and heterotopic transplantation experiments. Schematic representations of cell transplantations in embryos at the early gastrula stage (top), of late larval CNSs with schematic NB lineages (middle) and a table summarising the results (bottom). Numbers 1–4 indicate the transplantation modes and are reflected in the table below (T, thorax; A, abdomen). Only modes 1 and 4 (explantation from the thorax) give rise to large nests of postembryonic cells (with pNB; white part in schematic lineage; about two-thirds of cases) regardless of whether they develop in the thorax (1) or abdomen (4). Potential pNBs (?) in modes 2 and 3 (explantation from the abdomen) are rare and have very small lineages. Only a third of thoracically derived lineages (1, 4) but about 90% of abdominally derived lineages (2, 3) consist of embryonically born cells only (black in schematic lineage). nervous system there are dramatic differences between the numbers of pNBs present in the thoracic versus the abdominal neuromeres, where the homeotic genes Ubx and abd-A are expressed. The A1 and A2 neuromeres have an intermediate number of pNBs (Truman and Bate, 1988), and in the absence of Ubx the A1 neuromere contains more pNBs indicating that Ubx regulates the pattern of pNBs within the A1 neuromere (Truman et al., 1993). Our findings of supernumerary NBs in larvae heterozygous for Ubx suggest that UBX may also influence the behaviour of pNBs in more posterior neuromeres (similar transformations were observed in larval cuticles; Beachy, 1990). However, the main differences in the distribution and proliferation of pNBs in more posterior abdominal neuromeres are regulated by abd-A, as is evident from the aberrant pNB proliferation seen in neuromeres A3–A7 in Uab4 mutant larvae which is more reminiscent of the A2 segment and thus is analogous to transformations observed in the cuticle (Lewis, 1978). The regions where ectopic proliferation occurs when abd-A function is reduced (e.g. in larvae hemizygous for abd-A) correspond to those which have been reported to give rise to abnormal neuropil structures under the same conditions (Teugels and Ghysen, 1983; Ghysen et al., 1985; Ghysen and Lewis, 1986). Thus, these morphological abnormalities may result from supernumerary progeny of misregulated NBs. Homeotic regulation of NB proliferation has also been observed in the moth Manduca sexta (Booker and Truman, 1989; Miles and Booker, 1993) suggesting that this is a fundamental mechanism through which the development of the nervous system is regulated in insects and most likely in other organisms. The transition from the embryonic stages of neurogenesis to the postembryonic stages involves programmed cell death. This transition can be seen in stage 17 embryos when the number of grh-expressing cells in the abdomen decreases dramatically, and it is prevented by mutations that disrupt the cell-death pathway (Bray et al., 1989; White et al., 1994). In abd-A mutant embryos (Uab4) more grhexpressing cells persist in the abdomen, already reflecting the NB pattern which is later on detected by BrdU and toluidine blue in the larval nerve cord. Conversely, when ABD-A or UBX are ectopically expressed, fewer grhexpressing cells are found, suggesting that the homeotic genes are controlling whether or not these cells undergo programmed cell death. Interestingly, the numbers of abdominal NBs are also reduced upon misexpression of UBX or ABD-A (Fig. 2C vs. H,I). A possible explanation is that some NBs in A3–A7 (e.g. vm, vl and dl; Fig. 2A–C) never express functional amounts of UBX or ABD-A, so that targeted expression of UBX or ABD-A in those NBs could trigger their elimination. If so, then regulation of these NBs should normally be independent of abd-A and Ubx gene function. For example these NB could have a basal level of proliferation in all segments or they could undergo segment specific patterns of proliferation in response to other homeotic genes such as Antennapedia (see Fig. 1F). To distinguish these possibilities specific markers are needed for the individual NBs so that they can be identified in all segments and studied throughout embryonic and postembryonic stages. 3.2. Cellular memory is implicated in the regulation of homeotic gene expression and function The segment-specific effects of homeotic genes on proliferation and pNB fates are first detected quite late in embryogenesis, at about stage 15, and continue to be manifest throughout larval development (Truman and Bate, 1988; Bray et al., 1989; Prokop and Technau, 1991). We therefore expected that the late effect of homeotic mutations on NB behaviour would correlate with late expression of Ubx and abd-A in these cells. However, neither UBX nor ABD-A are present in NBs during stage 15–17 (Fig. 3D,E) nor during postembryonic stages (Fig. 3F; Truman et al., 1993), so they are unlikely to be directly controlling NB numbers or proliferation. However, ABD-A and UBX are present in many NBs at earlier stages and their presence is sufficient to induce abdominal regulation, as demonstrated by misexpression experiments. Likewise it has been shown A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 that the NB1-1 lineage requires Ubx and abd-A function during stage 10/11 of embryogenesis when the proteins can be detected in this precursor (Prokop and Technau, 1994a). Therefore, the activity of UBX and ABD-A at an early stage must be able to determine the later proliferation and persistence of the NBs (Fig. 6C). This is contrary to the original hypothesis, based on the requirements for Ubx in the epidermis (Morata and Garcı́a-Bellido, 1976), which proposed that the expression of a homeotic gene would need to be stably maintained in the progeny cells for them 107 to retain their identity. In the NBs there must be mechanisms downstream of homeotic genes through which cells can retain a stable memory or imprint of having previously expressed Ubx or abd-A. This could involve the homeotic genes acting as transcriptional repressors which initiate a repressed state of their target genes that can be maintained even after UBX or ABD-A have decayed. Alternatively, UBX or ABD-A may activate target genes which have the capacity for autoregulation, so that the targets can maintain their own expression in the absence of homeotic proteins. Fig. 6. Control of segment-specific proliferation and survival in Drosophila neuroblasts. The scheme shows (A) the expression domains of Antennapedia, Ubx and abd-A in relation to (B) the phenotypes observed (prolif., prolonged proliferation of embryonic NB; NB1-1, segmental differences of the NB1-1 lineage, see Prokop and Technau, 1994a; elimin., late embryonic NB elimination; no diff., possible NBs without segment specific differences, e.g. vm-, vl-, dl-pNB). (C) The implementation of genetic programmes for the regulation of NB numbers and proliferation depends on earlier patterning events. Mechanisms regulating persistence and prolonged proliferation of thoracic NB (+ in square) appear to operate after UBX or ABD-A have faded in NBs. Only those NBs expressing UBX (blue NBs) or ABD-A (red NBs) can prevent installation of these mechanisms (crossed out squares). Ubx (blue) and abd-A (red) are expressed only in those NBs which contain lineage or cell specific activators of homeotic transcription (e.g. segmentation genes; green NBs), and which have not expressed anterior gap genes (e.g. hunchback; NBs with grey gradients). The function of gap genes (gap, black) is maintained by Polycomb-group genes (Pc-group, grey) and both repress the NB-specific activation of Ubx or abd-A (bent green arrow). (D) Upon transplantation, precursors from the thorax (1) maintain their repressed state mediated by gap and Polycomb-group genes (grey gradient). Precursors from the abdomen (2) have not expressed inhibitory gap genes and can therefore express UBX and ABD-A (red/blue). 108 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 Although the activity of homeotic genes is required after NB segregation we found that at the early gastrula stage the neuroectodermal cells are already autonomously determined as to whether they will give rise to NBs with thoracic or abdominal proliferation behaviour, and whether they will persist as pNBs (Figs. 4 and 5). Gastrula cells transplanted from the thorax to the abdomen display thoracic patterns of proliferation, whereas cells transplanted from the abdomen to the thorax retain abdominal characteristics. This is likely to be the result of domain-specific repression of homeotic genes mediated by the early expressed gap gene products such as hunchback, a known repressor of Ubx (White and Lehmann, 1986). Once the proteins encoded by the gap genes have decayed, the repressed state of the homeotic genes is maintained by Polycomb group genes (Paro, 1990), so that the regulatory sequences of bithorax-complex genes in cells derived from the thorax remain repressed, whereas homeotic genes in abdominal cells are accessible to activation (Fig. 6C,D). Experiments in C. elegans have also demonstrated the importance of lineage rather than local positional signalling in the regulation of homeotic gene expression (Cowing and Kenyon, 1996), suggesting that there could be conservation of the mechanisms underlying this regulation. We conclude therefore that control of segment-specific development in the CNS involves a cascade of subsequent determination steps (Fig. 6C), in which first homeotic gene function is regionally permitted or not, and subsequently homeotic genes themselves select a programme, which can be maintained even after the bithorax-complex genes cease to be expressed. Reproducible intrasegmental expression patterns could spare some cells from homeotic gene expression (e.g. the abdominal vm-, vl- and dl-NB) and thus add further complexity and regulatory potential to this system. 4. Experimental procedures 4.1. Fly stocks We used the following mutant strains: AntennapediaW10 (Wakimoto et al., 1984); Ubx6.28 (Kerridge and Morata, 1982); abd-AMX1 (Sánchez-Herrero et al., 1985); the abd-A allele iab3Uab4 (Lewis, 1978, referred to as Uab4); Df(3R)Ubx109 (Lewis, 1978; referred to as Df109) and Pc3 (Lewis, 1978). For misexpression of homeotic genes in neural cells after NB segregation we used the GAL4 system (Brand and Perrimon, 1993) with the GAL4 driver-line MZ1407 (Sweeney et al., 1995; Urban et al., unpublished data) in conjunction with UAS-Ubx62.1 and UAS-abdA21.6 (Greig and Akam, 1993; Castelli-Gair et al., 1994). During misexpression experiments embryos were kept at 25 or 29°C. For cell transplantations we used as hosts the b-Gal-1 strain cq20 (Knipple and MacIntyre, 1984), which lacks endogenous b-galactosidase activity and therefore shows no background staining. As donors we used P-lacZ enhancer trap lines with strong expression of b-galactosidase throughout the nervous system. Either we used A45 (O’Kane and Gehring, 1987) or the line E;B1;Z1 carrying three different P-lacZ insertions: E is an insertion into the elav locus (Bier et al., 1989), B1 an unpublished insertion on the 2nd chromosome (Bier et al., 1989) and Z1 a 3rd chromosome insertion (unpublished data; kindly provided by J.A. Campos-Ortega). 4.2. Generation of transgenic flies containing grh-lacZ reporter gene A 4-kb fragment (XhoI-XbaI) from the second intron of the grainyhead gene (Uv et al., 1997) was cloned into pBluescript (Strategene), excised using XbaI and KpnI and inserted into the XbaI-KpnI sites of the enhancer detector P-element vector, HZ50PL (Hiromi et al., 1985) which carries rosy as a selectable marker. The resulting plasmid (NBgrh-lacZ) was injected into cinnabar rosy embryos in the presence of transposase using standard techniques (Rubin and Spradling, 1982) and transformants were selected on the basis of rosy + eyes. At least four independent lines were analysed, all of which had expression in NBs as shown. 4.3. Bromodeoxyuridine (BrdU) labelling For embryonic BrdU application 5–10 nl of a 15 mM solution of BrdU (Sigma) in 0.2 M KCl were injected into embryos about 45 min after the three-portioned midgut stage (stage16/17 according to Campos-Ortega and Hartenstein, 1997), following procedures described elsewhere (Prokop and Technau, 1991; Prokop and Technau, 1993). At late stage 17 embryos were mechanically removed from the vitelline membrane, washed in PBT, fixed in Carnoy’s solution (6:3:1 ethanol/chloroform/acetic acid), rehydrated and their tips cut off to allow subsequent penetration of antibodies. In larval stages BrdU was applied by feeding animals with standard medium containing between 1 and 10% of a BrdU-solution (33 mM BrdU in 40% ethanol; Truman and Bate, 1988). CNSs were removed, fixed in Carnoy’s solution, and rehydrated. To stain for BrdU-incorporation we used monoclonal antibodies against BrdU (Becton and Dickinson, diluted 1:100), HRP-coupled secondary antibodies (Dianova, diluted 1:500) or biotinylated secondary antibody (Dianova, diluted 1:500) followed by treatment with ABC elite kit (Vector Laboratories); blocking steps included 1% low fat milk powder. Specimens were embedded in Araldite (Serva) in borosilicate capillaries (Hilgenberg; see Prokop and Technau, 1993). 4.4. Immunocytochemistry Further antibody stainings were carried out on whole A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 mounts following standard procedures and using biotinylated secondary antibodies (1:300) and ABC elite kit (Vector Laboratories) or fluorescent secondary antibodies from Jackson laboratories (1:200). Third larval instar CNSs were briefly post fixed in methanol subsequent to formaldehyde fixation. For detection of grainyhead (grh) expression, the CNSs were dissected from stage 17 embryos (approximately 19–22 h after egg lay) and transferred to poly-l-lysine coated coverslips (Bray et al., 1989). Embryos and larval CNSs were embedded in Araldite (see above) or mounted in citifluor. The final dilutions of primary antibodies were 1:1500 for rat polyclonal anti-ABD-A (Macı́as et al., 1990); 1:20 for rabbit polyclonal anti-ABD-A (Karch et al., 1990), 1:20 for monoclonal anti-UBX (White and Wilcox, 1984, 1985), 1:3 for monoclonal anti-ENGRAILED (Patel et al., 1989); 1:5 for monoclonal anti-GRH (Bray et al., 1989); 1:10 000 for rabbit anti-ASENSE (Brand et al., 1993); and 1:2000 for rabbit anti-b-galactosidase (Cappell). 4.5. Cell transplantation, preparation and staining of donors and hosts For cell transplantations we used as donors enhancer trap lines with lacZ gene expression throughout the CNS (see Section 4.1). The lacZ gene is inherited upon cell division so that embryonic and postembryonic progeny of transplanted cells produce b-galactosidase cell autonomously. In addition, donors were injected with 5–10 nl of horseradish peroxidase solution (HRP; 5–10% in 0.2 M KCl) at the syncytial blastoderm stage. Transplantations of single cells were carried out about 10 min after the onset of gastrulation (stage 7) in the ventral neurogenic region at either 30% (abdominal) or at 60% egg length (thoracic, Fig. 5; for more details see Prokop and Technau, 1991). Late larval host CNSs were dissected out in PBS (other tissues were discarded), fixed in 1% solution of glutaraldehyde in PBS, stained for b-galactosidase and subsequently for HRP activity. Specimens were dehydrated in alcohol, cleared in xylene, embedded in araldite (see above). Acknowledgements We thank José Campos Ortega, James Castelli-Gair, Stephen Greig, Rolf Reuter, Srikala Raghavan, Suzanna Romani, Ernesto Sánchez Herrero, Joachim Urban and Rob White for providing fly stocks and antibodies, Rob White and Ernesto Sánchez-Herrero for comments on the manuscript. A.P. is grateful to Michael Bate, in whose laboratory part of the project was carried out. This work was supported by a grant from the Deutsche Forschungsgemeinschaft to G.M.T. (Te 130/1-4), by a project grant from the Biotechnology and Biological Science Research Council to S.J.B. and by a Human Capital and Mobility Fellowship (EU) and a research fellowship from the Lloyd’s of London Tercentenery Foundation to A.P. 109 References 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. Morphol. 35, 107–123. Beachy, P.A., 1990. A molecular view of the Ultrabithorax homeotic gene of Drosophila. Trends Genet. 6, 46–51. Bier, E., Vässin, H., Shepherd, S., Lee, K., McCall, K., Barbel, S., Ackerman, L., Carretto, R., Uemura, T., Grell, E., Jan, L.Y., Jan, Y.N., 1989. Searching for pattern and mutation in the Drosophila genome with a P-lac Z vector. Genes Dev. 3, 1273–1287. Booker, R., Truman, J.W., 1989. Octopod, a homeotic mutation of the moth Manduca sexta, influences the fate of identifiable pattern elements within the CNS. Development 105, 621–628. Bossing, T., Technau, G.M., 1994. The fate of the CNS midline progenitors in Drosophila as revealed by a new method for single cell labelling. Development 120, 1895–1906. Bossing, T., Udolph, G., Doe, C.Q., Technau, G.M., 1996. The embryonic central nervous system lineages of Drosophila melanogaster. I. The neuroblast lineages derived from the ventral half of the neuroectoderm. Dev. Biol. 179, 41–64. Brand, A.H., Perrimon, N., 1993. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401–415. Brand, M., Jarman, A.P., Jan, L.Y., Jan, Y.N., 1993. asense is a Drosophila neural precursor gene and is capable of initiating sense organ formation. Development 119, 1–17. Bray, S.J., Burke, B., Brown, N.H., Hirsh, J., 1989. Embryonic expression pattern of a family of Drosophila proteins that interact with a central nervous system regulatory element. Genes Dev. 3, 1130–1145. Campos-Ortega, J.A., Hartenstein, V., 1997. The Embryonic Development of Drosophila melanogaster. Springer–Verlag, Berlin, 405 pp. Castelli-Gair, J., Greig, S., Micklem, G., Akam, M., 1994. Dissecting the temporal requirements for homeotic gene function. Development 120, 1983–1995. Cowing, D., Kenyon, C., 1996. Correct Hox gene expression established independently of position in Caenorhabditis elegans. Nature 382, 353– 356. Doe, C.Q., 1992. Molecular markers for identified neuroblasts and ganglion mother cells in the Drosophila central nervous system. Development 116, 855–886. Doe, C.Q., Hiromi, Y., Gehring, W.J., Goodman, C.S., 1988. Expression and function of the segmentation gene fushi tarazu during Drosophila neurogenesis. Science 239, 170–175. Doe, C.Q., Scott, M.P., 1988. Segmentation and homeotic gene function in the developing nervous system of Drosophila. Trends Neurosci. 11, 101–106. Duncan, I., 1987. The bithorax complex. Annu. Rev. Genet. 21, 285–319. Ghysen, A., Jan, L.Y., Jan, Y.N., 1985. Segmental determination in Drosophila central nervous system. Cell 40, 943–948. Ghysen, A., Lewis, E.B., 1986. The function of bithorax genes in the abdominal central nervous system of Drosophila. Roux’s Arch. Dev. Biol. 195, 203–209. González-Reyes, A., Morata, G., 1990. The developmental effect of overexpressing a Ubx product in Drosophila embryos is dependent on interactions with other homeotic products. Cell 61, 515–522. Greig, S., Akam, M., 1993. Homeotic genes autonomously specify one aspect of pattern in the Drosophila mesoderm. Nature 362, 630– 632. Hartenstein, V., Campos-Ortega, J.A., 1984. Early neurogenesis in wild type Drosophila melanogaster. Roux’s Arch. Dev. Biol. 193, 308–325. Hiromi, Y., Kuroiwa, A., Gehring, W.J., 1985. Control elements of the Drosophila segmentation gene fushi tarazu. Cell 43, 603–613. Karch, F., Bender, W., Weiffenbach, B., 1990. abd-A expression in Drosophila embryos. Genes Dev. 4, 1573–1587. Kaufman, T.C., Seeger, M.A., Olsen, G., 1990. Molecular and genetic 110 A. Prokop et al. / Mechanisms of Development 74 (1998) 99–110 organization of the Antennapedia gene complex of Drosophila melanogaster. Adv. Genet. 27, 309–362. Kerridge, S., Morata, G., 1982. Developmental effects of some newly induced Ultrabithorax alleles of Drosophila. J. Embryol. Exp. Morphol. 68, 211–234. Knipple, D.C., MacIntyre, R.J., 1984. Cytogenic mapping and isolation of mutations of the b-gal-1 locus of Drosophila melanogaster. Mol. Gen. Genet. 198, 75–83. Lewis, E.B., 1978. A gene complex controlling segmentation in Drosophila. Nature 276, 565–570. Macı́as, A., Casanova, J., Morata, G., 1990. Expression and regulation of the abd-A gene of Drosophila. Development 110, 1197–1207. McGinnis, W., Krumlauf, R., 1992. Homeobox genes and axial patterning. Cell 68, 283–302. Miles, C.I., Booker, R., 1993. Octopod, a homeotic mutation of the moth Manduca sexta, affects development of both mesodermal and ectodermal structures. Dev. Biol. 155, 147–160. Morata, G., 1993. Homeotic genes of Drosophila. Curr. Opin. Genet. Dev. 3, 606–614. Morata, G., Garcı́a-Bellido, A., 1976. Developmental analysis of some mutants of the bithorax system of Drosophila. Roux’s Arch. Dev. Biol. 179, 125–143. O’Kane, C., Gehring, W.J., 1987. Detection of in situ genomic regulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA 84, 9123–9127. Paro, R., 1990. Imprinting a determined state into the chromatin of Drosophila. Trends Genet. 6, 416–421. Patel, N.H., Martin-Blanco, E., Coleman, K.G., Poole, S.J., Ellis, M.C., Kornberg, T.B., Goodman, C.S., 1989. Expression of engrailed proteins in arthropods, annelids and chordates. Cell 58, 955–968. Prokop, A., Technau, G.M., 1991. The origin of postembryonic neuroblasts in the ventral nerve cord of Drosophila melanogaster. Development 111, 79–88. Prokop, A., Technau, G.M., 1993. Cell transplantation. In: Hartley, D. (Ed.), Cellular Interactions in Development: A Practical Approach. Oxford University Press, Oxford, pp. 33–57. Prokop, A., Technau, G.M., 1994a. Early tagma-specific commitment of Drosophila CNS progenitor NB1-1. Development 120, 2567–2578. Prokop, A., Technau, G.M., 1994b. Normal function of the mushroom body defect gene of Drosophila is required for the regulation of the number and proliferation of neuroblasts. Dev. Biol. 161, 321–337. Prokop, A., Technau, G., 1995. BrdU incorporation reveals DNA replication in non dividing glial cells in the larval abdominal CNS of Drosophila. Roux’s Arch. Dev. Biol. 204, 54–61. Rubin, G.M., Spradling, A.C., 1982. Genetic transformation of Drosophila with transposable element vectors. Science 218, 348–353. Sánchez-Herrero, E., Vernos, I., Marco, R., Morata, G., 1985. Genetic organization of Drosophila bithorax-complex. Nature 313, 108–113. Schmidt, H., Rickert, C., Bossing, T., Vef, O., Urban, J., Technau, G.M., 1997. The embryonic central nervous system lineages of Drosophila melanogaster. II. Neuroblast lineages derived from the dorsal part of the neuroectoderm. Dev. Biol. 189, 186–204. Simon, J., Chiang, A., Bender, W., 1992. Ten different Polycomb group genes are required for spatial control of the abd-A and Abd-B homeotic products. Development 114, 493–505. Sweeney, S.T., Broadie, K., Keane, J., Niemann, H., O’Kane, C.J., 1995. Targeted expression of tetanus toxin light chain in Drosophila specifically eliminates synaptic transmission and causes behavioural defects. Neuron 14, 341–351. Teugels, E., Ghysen, A., 1983. Independence of the number of legs and leg ganglia in Drosophila bithorax mutants. Nature 304, 440–442. Truman, J.W., Bate, C.M., 1988. Spatial and temporal patterns of neurogenesis in the CNS of Drosophila melanogaster. Dev. Biol. 125, 145– 157. Truman, J.W., Taylor, B.J., Awad, T.A., 1993. Formation of the adult nervous system. In: Bate, M., Martı́nez Arias, A. (Eds.), The development of Drosophila melanogaster. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 1245–1275. Udolph, G., Prokop, A., Bossing, T., Technau, G.M., 1993. A common precursor for glia and neurons in the embryonic CNS of Drosophila gives rise to segment-specific lineage variants. Development 118, 765–775. Uv, A.E., Harrison, E.J., Bray, S.J., 1997. Tissue-specific splicing and functions of the Drosophila transcription factor Grainyhead. Mol. Cell Biol. 17, 6727–6735. Wakimoto, B.T., Turner, F.R., Kaufmann, T.C., 1984. Defects in embryogenesis in mutants associated with the Antennapedia gene complex of Drosophila melanogaster. Dev. Biol. 102, 147–172. White, K., Grether, M.E., Abrams, J.M., Young, L., Farrell, K., Steller, H., 1994. Genetic control of programmed cell death in Drosophila. Science 264, 677–683. White, R.A.H., Lehmann, R., 1986. A gap gene, hunchback, regulates the spatial expression of Ultrabithorax. Cell 47, 311–321. White, R.A.H., Wilcox, M., 1984. Protein products of the bithorax complex in Drosophila. Cell 39, 63–171. White, R.A.H., Wilcox, M., 1985. Distribution of Ultrabithorax proteins in Drosophila. EMBO J. 4, 2035–2043.