Developmental defects in brain segmentation caused by mutations of the homeobox genes orthodenticle and empty spiracles in Drosophila

Neuron, Vol. 15, 769-778, October, 1995, Copyright © 1995 by Cell Press

Developmental Defects in Brain Segmentation Caused by Mutations of the Homeobox Genes orthodenticle and empty spiracles in Drosophila Frank Hirth,* Stavros Therianos,* Thomas Loop,* Walter J. Gehring,t Heinrich Reichert,* and Katsuo Furukubo-Tokunaga*t *Laboratory of Neurobiology Institute of Zoology University of Basel CH-4051 Basel Switzerland tDepartment of Cell Biology Biocenter University of Basel CH-4056 Basel Switzerland

Summary We have studied the roles of the homeobox genes orthodenticle (otd) and empty spiracles (ems) in embryonic brain development of Drosophila. The embryonic brain is composed of three segmental neuromeres. The otd gene is expressed predominantly in the anterior neuromere; expression of eros is restricted to the two posterior neuromeres. Mutation of old eliminates the first (protocerebral) brain neuromere. Mutation of eros eliminates the second (deutocerebral) and third (tritocerebral) neuromeres, otd is also necessary for development of the dorsal protocerebrum of the adult brain. We conclude that these homeobox genes are required for the development of specific brain segments in Drosophila, and that the regionalized expression of their homologs in vertebrate brains suggests an evolutionarily conserved program for brain development. Introduction The question of how the development of the brain is genetically controlled is one of the central issues of neural science. During development, an enormous number of neurons are generated in complex arrangements and subsequently become interconnected into precise neuronal circuits. The genetic and molecular mechanisms that underlie these biological processes are largely unknown. Neuroanatomical studies in vertebrates indicate that the developing brain may be organized into distinct neuromeric regions that reflect a basic metameric organization (reviewed in Guthrie, 1995). Similarly, morphological and developmental studies on the insect brain also suggest an underlying metameric organization. The insect brain is thought to derive from the neuromeres corresponding to either three or four modified body segments; however, there is considerable debate over the exact nature of seg~:Presentaddress: Insituteof BiologicalSciences,TsukubaUniversity, Tsukuba 305, Japan.

mentation in the insect head and brain (see Rempel, 1975; Cohen and JL~rgens, 1991; Finkelstein and Boncinelli, 1994). In vertebrates, recent developmental studies of the hindbrain reveal a clear metameric organization based on rhombomeres with distinct anatomical cytoarchitectures (see Lumsden, 1990). Moreover, molecular studies show that the expression patterns of Hox genes are segmentally restricted in these rhombomeres, supporting the notion that the rhombomeres are the segmental units in the hindbrain. Hox gene expression in these rhombomeres occurs in overlapping domains along the anterior-posterior axis (reviewed in Keynes and Krumlauf, 1994). Inactivation of individual Hox genes by targeted disruption results in respecification of the rhombomeres, comparable to what is caused by homeotic transformation in the trunk segments of Drosophila melanogaster. Although they are expressed in many parts of the vertebrate CNS, genes of the Hox clusters are not expressed in the forebrain and midbrain. These complex anterior regions of the developing brain are characterized by the regionalized expression of other putative regulatory genes; comparative studies of the restricted expression domains of these genes form the basis for a neuromeric model of the vertebrate forebrain (Puelles and Rubenstein, 1993; Rubenstein et al., 1994). Among the developmental regulatory genes with temporally and spatially restricted expression patterns in the vertebrate forebrain are the genes Otx and Emx. These are the vertebrate homologs of the genes orthodenticle (otd) and empty spiracles (ems) originally identified in Drosophila. Based on homology between the hom eobox sequences, Otx and Emx genes have now been isolated in various vertebrates including zebrafish, mouse, and human (Simeone et al., 1992a, 1992b, 1993; reviewed in Finkelstein and Boncinelti, 1994). In Drosophila, otd and ems are expressed in overlapping domains at the anterior pole of the cellular blastoderm stage embryo under the control of the bicoidgene product (Dalton et al., 1989; Cohen and J0rgens, 1990; Finkelstein and Perrimon, 1990; Finkelstein et al., 1990; Walldorf and Gehring, 1992). Mutational inactivation of these genes results in specific defects in epidermal head segments. Given the postulated role of the homologs of these genes in vertebrate brain development, it becomes important both in terms of development and in terms of evolution to determine whether the homeobox genes otd and ems are involved in embryonic brain development in Drosophila. In this paper we use high resolution molecular neuroanatomical techniques to show that both otd and eros genes are expressed in the developing Drosophila brain in segmental domains delimited by the segmentation gene engrailed(en). We further demonstrate that mutational inactivation of otd and ems genes results in distinct segmental defects in the developing fly brain, otd gene activity is required for the formation of the protocerebrum anlage; eros gene activity is required for the formation of the deu-

Neuron 77O

A

~i ~

~!i

Figure 1. Expression of en in the Embryonic Brain Laser confocal microscopy of wild-type stage 12 (A-D) and stage 15 (E and F) embryos double labeled with anti-en antibody (yellow/green) and neuron-specific anti-HRP antibody (red/orange; A, B, D, and F) or labeled with neuron-specific anti-HRP antibody (red/orange; C and E). Images are reconstructions of optical sections. (A) The two most anterior stripes of en-expressing cells are visible in this frontal view of the embryonic brain. The en-bl stripe is labeled with arrows, the en-b2 stripe with arrowheads. The en-b3 stripe is not visible in this focal plane. (B) The posterior stripe of en-expressing cells in the embryonic brain, en-b3, is indicated with open arrows. Sections were taken at a different focal plane than in (A). (C) Lateral view of embryonic brain. Only neuron-specific labeling with the anti-HRP antibody is shown for comparison with (D). (D) Lateral view of embryonic brain. Same preparation as in (C) but with both en and neuron-specific anti-HRP labeling shown, en-expressing cells in the brain appear as spots or clusters in the laterat view. Arrow, en-bl; arrowhead, en-b2; open arrow, en-b3. (E) Frontal view of embryonic brain. Only neuron-specific labeling with the anti-HRP antibody is shown for comparison with (F). (F) Frontal view of embryonic brain. Same preparation as in (E) but with both en and neuron-specific anti-HRP labeling shown. The two most anterior stripes of en-expressing cells in the brain are still detectable in this frontal view. en-bl is indicated with arrows, en-b2 with arrowheads; en-b3 is not visible in this focal plane. The small groups of en-expressing cells anterior to en-bl derive from the en-bl stripe and move anteriorly ("secondary head spot" from Schmidt-Ott and Technau, 1992). A single en-expressing cell is seen at the midline of the preoral brain commissure. Bar, 10 I~m. tocerebrum and tritocerebrum aniagen. Our results indicate that the o t d and e m s h o m e o b o x genes have a central regulatory role in brain morphogenesis and regionalization during embryonic d e v e l o p m e n t in Drosophila. Moreover, the regionalized brain expression domains of otd/Otx and e m s / E m x g e n e s in flies and vertebrates imply that genetic regulatory mechanisms for brain patterning are highly conserved.

Results The Expression Pattern of the e n Gene Reveals a Subdivision of the Embryonic Brain into Three Neuromeres The cephalic region of the Drosophila e m b r y o is composed of diverse anatomical structures that undergo a series of complex morphogenetic m o v e m e n t s as head involution occurs (Campos-Ortega and Hartenstein, 1985; J~rgens and Hartenstein, 1993). To identify the anlagen of the era-

bryonic brain in this c o m p l e x cephalic region, we carried out high resolution confocal microscopy coupled with imm u n o c y t o c h e m i c a l double labeling techniques. As a molecular l a n d m a r k for the segmental subdivisions of the embryonic brain, we focused on the expression pattern of the segment polarity gene en (see Diederich et al., 1991; Schmidt-Ott and Technau, 1992; Schmidt-Ott et al., 1994). As a marker for neuronal cells in the brain, we used the neuron-specific anti-horseradish peroxidase (HRP) antibody (Jan and Jan, 1982). (For details on the neuroanato m y of the Drosophila embryonic brain, see Therianos et al., 1995.) Between stages 11 and 12 of embryogenesis, three groups of en-expressing cells are observed in each hemisphere of the embryonic brain (Figures 1A and 1B). We refer to these stripes of en-expressing cells in the brain as e n - b l , en-b2, and en-b3 to distinguish them from adjacent epidermal en-expressing cells, which have different developmental fates.

Brain Developmentin Drosophila 771

The most anterior (according to neuraxis) group of enexpressing cells, en-bl, forms a horizontal stripe that extends from the outer surface to the medial edge of each hemisphere at stage 12 (Figure 1A). Viewed from a lateral perspective, the en-bl stripe appears as a single cluster (Figures 1C and 1D); in this lateral view, the en-bl stripe corresponds to the previously described "en head spot" (Cohen and JL~rgens, 1990; Schmidt-Ott and Technau, 1992). As embryogenesis proceeds, these cells continue to express en, albeit at a lower level. At embryonic stage 15, en-bl cells can be detected as broken stripes located in each hemisphere just posterior to the differentiated preoral embryonic brain commissure (Figures 1E and 1F). The second stripe of en-expressing cells in the brain, en-b2, is located more posteriorly along the neuraxis. At stage 12, it extends from the lateral edge of each hemisphere medially through the brain toward the frontal nerve commissure (Figure 1B). Although its elongated linear morphology is clear in a frontal view, from a lateral view the en-b2 stripe appears as a single cluster of cells (Figures 1C and 1D) and may, therefore, correspond to the "en antennal spot" (Schmidt-Ott and Technau, 1992). The enb2 stripe, like the en-b I stripe, can be observed throughout the rest of embryogenesis. At stage 15, the en-b2 cells can be recognized as broken stripes crossing the embryonic brain at the level of the developing antennal nerve (Figures 1 E and 1 F). The en-b3 stripe is located in the most posterior (according to neuraxis) part of the brain, at the juncture between the brain and the subesophageal ganglion (Figures 1B and 1D). The cells of the en-b3 stripe persist in later embryonic stages. At stage 15, the en-b3 cells are still detectable as a stripe located at the level of the tritocerebral commisssure. The characteristic anatomical location of the en-b3 stripe suggests that it may correspond to the en stripe of the "intercalary neuromere" described previously (Diederich et al., 1991). The existence of three distinct stripes of en-expressing cells in the embryonic brain argues for the metameric division of the embryonic Drosophila brain into three neuromeres (see Figure 7). We refer to these as neuromere bl, neuromere b2, and neuromere b3. We posit that these three neuromeres correspond to the aniagen of the protocerebrum, deutocerebrum, and tritocerebrum of the mature brain.

The otd and eros Homeobox Genes Are Expressed in Specific Neuromeres of the Embryonic Brain The otd and ems homeobox genes are known to act as anterior gap genes during Drosophila embryogenesis (J~rgens et al., 1984; Dalton et al., 1989; Cohen and J~rgens, 1990, 1991; Walldorf and Gehring, 1992; Finketstein and Boncinelli, 1994). The activity of the otd and ems genes is required in the development of the epidermal head segments, including antennal and preantennal structures. The otdgene is also required in ventral CNS development; mutations in otd result in a malformation of the neuropil, the disappearance of midline-associated neurons, and fused commissures in the ventral nerve cord (Finkelstein et al., 1990; Kl&mbt et al., 1991). The expression of ems

in the embryonic CNS has been described briefly as occurring in small clusters of lateral neuroblasts in the gnathai lobes and in several cells per segment at the lateral margins of the ventral nerve cord (Dalton et al., 1989; Walldorf and Gehring, 1992). To determine the expression domains of otd and ems in the embryonic brain, we carried out immunocytochemical double labeling experiments with an anti-otd antibody (Wieschaus et al., 1992) or an anti-ems antibody (Walldorf and Gehring, 1992) in conjunction with a neuron-specific anti-HRP antibody (Jan and Jan, 1982) or an anti-en antibody (Patel et al., 1989). At embryonic stage 12, otd immunoreactivity is seen throughout most of the neuronal cell body regions of brain neuromere b l , with the exception of the most anterior (according to neuraxis) regions (Figure 2A). otd immunoreactivity is also seen at the border between neuromeres bl and b2 and overlaps partially with the anterior part of neuromere b2. At stage 16, otd expression is similar in its spatial extent (Figures 2B and 2C). Most of the cell body regions in neuromere bl show otd immunoreactivity, as do cells at the boundary region between neuromeres bl and b2 and in the most anterior part of neuromere b2. The parts of neuromere bl that are not immunolabeled are the most anterior (according to neuraxis) regions, as well as the neuropil, axon tracts, and commissure (note that the otd protein is a nuclear factor). Similar consistent patterns were observed by in situ hybridization for the otdtranscript using otd cDNA (Finkelstein et al., 1990) as a probe (data not shown). Immunostaining of glial cells along the commissure and at the medial edges of the brain tracts is probably due to cross-reactivity of the anti-otd antibody with the homeodomain of the glial-specific reversed polarity (repo)/RK2 protein (Campbell et al., 1994; Xiong et al., 1994; Halter et al., 1995). Expression of ems in the embryonic brain is found in two neural cell body groups located in neuromeres b2 and b3 (Figure 3A). The larger, more anterior (according to neuraxis) ems-immunoreactive cell body group is located in neuromere b2 and extends from the en-bl stripe posteriorly to the en-b2 stripe (Figures 3B and 3C). With the exception of some of the cells that make up the en-bl and en-b2 stripes at the neuromere's borders (Figure 3C), most of the neuronal cell bodies in neuromere b2 appear to be ems-immunoreactive. The smaller, more posterior emsimmunoreactive cell body cluster is found in neuromere b3 and extends from the en-b2 stripe to the tritocerebral commissure (Figures 3B and 3C). As in neuromere b2, most of the neuronal cell bodies in neuromere b3 appear to be ems immunoreactive. Additional ems immunoreactivity is detected in a few glial cells associated with the preoral brain commissure and the longitudinal connectives. Neuropil and tracts in neuromeres b2 and b3 are not immunolabeled, consistent with the nature of the eros protein as a nuclear factor, ems expression is seen in segmentally repeated cell clusters in the neuromeres of the subesophageal ganglion (Figure 3A) as previously reported (Dalton et al., 1989; Walldorf and Gehring, 1992). In summary, the otd gene is expressed in neural cell body regions throughout much of brain neuromere bl, as well as in the border region between neuromeres bl and

Neuron 772

C

Figure 2. Expression of otd in the Embryonic Brain Laser confocal microscopy of wild-type stage 12 (A) and stage 16 (B and C) embryos double labeled with anti-otd antibody (yellow/green) and neuron-specific anti-HRP antibody (red/orange). Images are reconstructions of optical sections. (A) Lateral view o f o t d expression in the embryonic brain. Arrowheads, otd-expressing regions; open arrows, epidermal otd staining outside of the brain. (B) Frontal view of otd expression in the embryonic brain. Otd-expressing regions are delimited by arrowheads. Open arrow indicates immunostaining at the preoral brain commissure due to cross-reactivity of the anti-otd antibody with the glial-specific repo/RK2 protein. (C) Frontal view of otd expression in the embryonic brain. Same preparation as in (8), but sections were taken at a different focal plane, otd-expressing regions are delimited by arrowheads. Open arrows indicate immunostaining at the medial edges of the descending brain tracts due to cross-reactivity as in (B). Bar, 10 p.m.

b2. The e m s gene is expressed in many of the neural cell body regions in brain neuromeres b2 and b3. These expression patterns suggest that the o t d h o m e o b o x gene might be important for the development of brain neuromere b l and that the e m s h o m e o b o x gene might be important for the development of brain neuromeres b2 and b3.

Mutations in the otd and e m s Homeobox Genes Affect the Formation of Specific Brain Neuromeres Mutations in the genes o t d and e r o s are known to result in the malformation of the epidermal cuticular patterns

in embryonic head (see above). What are the effects of mutations in these genes on the development of the embryonic brain? To address this question, we used high resolution confocal microscopy techniques to analyze the embryonic brain phenotypes of several different alleles (including nulls) of these genes. A brief summary of these alleles and of the observed brain phenotypes is given in Tables 1 and 2. Consistent with its neuromere-specific expression pattern, mutations in the o t d g e n e result in absence or severe reduction of neuromere b l of the embryonic brain (Figure 4A). Thus, in all the alleles of o t d examined, more than

Figure 3. Expression of ems in the Embryonic Brain Laser confocal microscopy of wild-type stage 12 (A and C) and stage 14 (B) embryos double labeled with anti-ems antibody (yellow/green) and neuron-specific anti-HRP antibody (red/orange; A and B) or anti-en antibody (red/orange; C). Images are reconstructions of optical sections. (A) Lateral view of ems expression in the embryonic brain. Arrows, ems expression in neuromere b2; arrowheads, eros expression in neuromere b3. (B) Frontal view of ems expression in the embryonic brain. Arrow, ems expression in neuromere b2; arrowhead, eros expression in neuromere b3. (C) Frontal view of ems expression and en expression in the embryonic brain. Different preparation and different focal plane of optical sections than in (B). Arrows, ems expression in neuromere b2; arrowheads, ems expression in neuromere b3; open arrows, en expression in en-b2 and en-b3 stripes. Bar, 10 ~m.

Brain Development in Drosophila 773

Table 1. Summary of Segmental Brain Defects in orthodenticle Alleles Brain Defects Alleles

+++

++

+

otd JAI01 otd YH13 otdXD87

72 (80.8%) 14 (15.7%) 3 (3.3%) 115 (90.0%) 10 (7.8%) 2 (1.5%) 120 (80.0%) 16 (10.6%) 14 (9.3%)

Total Embryos Examined 89 127 150

Brain defects were separated into three phenotypes: neuromere bl, including the brain commissure, is absent, and sometimes embryonic brain is exposed from the ectoderm (+++); most of neuromere bl is absent and the brain commissure is severely reduced (++); a remnant of neuromere bl is observed with reduced brain commissure (+). JAI01 is protein null due to a small deletion in the coding region (Finkelstein et al., 1990). YH13 and XD87are genetically null for epidermal traits (Finkelstein et al., 1990).

Table 2. Summary of Segmental Brain Defects in empty spiracles Alleles Brain Defects Alleles

+++

++

ems 9H83 116 (87.2%) 9 (6.7%) eros 10D29 103 (88.0%) 7 (5.9%)

+

Total Embryos Examined

8 (6.0%) 7 (5.9%)

133 t17

Brain defects were separated into three phenotypes: neuromeres b2 and b3 are absent, and sometimes embryonic brain is exposed from the ectoderm [+++), neuromeres b2 and b3 are severely reduced and deformed (++); neuromeres b2 and b3 are only slightly affected (+). 9H83 is protein null with an aberrant stop codon in the coding region (Dalton et al., 1989). The molecular lesion of 10D29 (Jurgens et al., 1984) is unknown.

8 0 % of the mutant embryos fail to develop neural structures anterior to the e n - b l stripe (Table 1). In these embryos, the preoral brain commissure is completely absent, resulting in two separate rest hemispheres composed of neuromeres b2 and b3 (Figure 4C). Mutations in the o t d gene have no gross morphological effects on the formation of brain neuromeres b2 and b3, both of which appear to differentiate normally (Figure 4A). Thus, the e n - b l stripe is present in most mutant embryos (data not shown). The expression pattern of the e m s gene in neuromeres b2 and b3 appears normal (Figure 4B). The circumesophageal connectives that interconnect the ventral part of the brain and the ventral nerve cord are formed and are closely associated with repo-expressing longitudinal glial cells (Figure 4D). (For a description of cirumesophageal connectives development, see Theftanos et at., 1995.) Mutation of the e m s gene leads to strong diminuation or complete absence of neuromeres b2 and b3 of the embryonic brain. In nearly 9 0 % of the mutant phenotypes, a clear gap between brain neuromere bl and the first (mandibular) neuromere of the subesophageat ganglion is observed (Figure 5A; Table 2). In these mutants, the e n - b 2 and e n - b 3 stripes are missing (Figure 5B). Neither the antennal nerve nor the tritocerebral commissure is retained. In contrast, mutational inactivation of the e m s gene has no gross morphological effect on the formation of brain neuromere b l (Figure 5A). The e n - b l stripe is still detectable at the appropriate position in the neuromere (Figure Figure 4. The Embryonic Brain in otd Mutants Laser confocal microscopy of otd mutant stage 12 (A and B) and stage 13 (C and D) embryos labeled with neuron-specificanti-HRP antibody (red/orange; A and C), double labeled with antiems antibody (yellow/green) and neuronspecific anti-HRP antibody (red/orange; B), or double labeled with anti-repo/RK2 antibody (yellow/green) and neuron-specific anti-HRP antibody (red/orange; D). Images are reconstructions of optical sections. (A) Lateral view of mutant embryonic brain. Only neuron-specific labeling with the antiHRP antibody is shown for comparison with (B). Asterisk, the missing brain neuromere bl. (B) Lateral view of mutant embryonic brain. Same preparation as in (A) but with both ems and neuron-specific anti-HRP labeling shown. Arrow, eros expression in neuromere b2; arrowhead, ems expression in neuromere b3. (C) Frontal view of mutant embryonic brain. Only neuron-sPecific labeling with the antiHRP antibody is shown for comparison with (D). The frontal nerve commissure of the stomatogastric nervous system is still present (open arrow). Arrowhead, missing preoral brain commissure; asterisks, missing brain neuromere bl. (D) Frontal view of mutant embryonic brain. Same preparation as in (C) but with both antirepo/RK2 and neuron-specific anti-HRP labeling shown. Open arrows, longitudinal glial celts associated with the embryonic circumesophageal connectives; asterisks, missing brain neuromere bl. Bar, 10 #.m.

Neuron 774

Figure 5. The Embryonic Brain in ems Mutants Laser confocal microscopy of eros mutant stage 12 (A and B) and stage 13 (C and D) embryos labeled with neuron-specific antiHRP antibody (red/orange; A and C), double labeled with anti-en antibody (yellow/green) and neuron-specific anti-HRP antibody (red/orange; B), and double labeled with anti-repo/ RK2 antibody (yellow/green) and neuronspecific anti-HRP antibody (red/orange; D). Images are reconstructions of optical sections. (A) Lateral view of mutant embryonic brain. Only neuron-specific labeling with the antiHRP antibody is shown for comparison with (B). Arrow, missing brain neuromeres b2 and b3. (B) Lateral view of mutant embryonic brain. Same preparation as in (A) but with both en and neuron-specific anti-HRP labeling shown. Arrow indicates the en-bl stripe viewed from lateral perspective that is still present in neuromere bl. (C) Frontal view of mutant embryonic brain. Only neuron-specific labeling with the antiHRP antibody is shown for comparison with (D). Arrowhead, preoral brain commissure; asterisks, missing brain neuromeres b2 and b3. (D) Frontal view of mutant embryonic brain. Same preparation as in (C) but with both antirepo/RK2 and neuron-specific anti-HRP labeling shown. Open arrows, midline glial cells associated with the preoral embryonic brain commissure; asterisks, missing brain neuromeres b2 and b3. Bar, 10 I~m.

5B). A preoral embryonic brain commissure is formed between the two hemispheres of neuromere b l (Figure 5C), and this commissure is associated with repo-expressing midline glial cells (Figure 5D). (For a description of brain commissure development, see Therianos et al., 1995.) In summary, we conclude that mutations of the homeobox genes o t d and eros have dramatic and specific effects on segmental development in the embryonic brain. Mutational inactivation of the o t c / g e n e leads to elimination or severe reduction of brain neuromere b l . Mutational inactivation of the eros gene leads to elimination or severe reduction of brain neuromeres b2 and b3.

The otd Gene Is Required for the Development of the Protocerebral Bridge in the Adult Brain The embryonic brain continues to develop during the larval stages and undergoes extensive remodeling during the pupal stage to establish the complex adult brain (Hartenstein, 1993; Truman et al., 1993). During this time, o t d activity is necessary for the specification of medial cell fates in both the larval and adult epidermis, including the formation of ocelti and ocellar bristles on the adult head (Wieschaus et al., 1984). In situ hybridization for o t d t r a n s c r i p t s reveals a regionalized expression in the prepupal brain, with strong expression detected in the medial cells in the ventral nerve cord and in several subregions in the central brain, including the dorsomedial regions (data not shown). To investigate possible functions of this postembryonic expression of o t d

in the central brain, we histologically examined the brains of a viable o t d allele, ocelliless (oc). Our analysis shows clear defects in the formation of the protocerebral bridge located atthe dorsomedial region of the adult brain (Figure 6). The protocerebral bridge is part of the central complex and is required for locomotor coordination in adult flies (Strauss and Heisenberg, 1993). The observed anatomical defects were seen in all brains examined for the alleles o c ~ and o c ~a~ but not in oc db, the weakest allele.

Discussion The findings reported here have structural, functional, and evolutionary implications for understanding brain development. The expression of en in stripes of cells at the borders of the three brain neuromeres and the regionalized expression patterns of the o t d and ems homeobox genes in these neuromeres indicate a metameric organization of the embryonic Drosophila brain (Figure 7). The fact that mutational inactivation of o t d and e m s genes results in dramatic defects in brain neuromere morphogenesis implies that the early regionalization of the Drosophila brain is controlled by these homeobox genes. Furthermore, since the homologs of these genes, O t x and Emx, are expressed in a regionalized manner in the vertebrate brain, their involvement in brain development may be evolutionarily conserved and general in all higher brains. We discuss the major implications of these findings below.

8rain Developmentin Drosophila 775

Figure 6. otd is Requiredfor Developmentof the Mature Brain Frontal 7 ~m sectionsof the brains of adult wild-type(A) and ocelliless, In(1)z~49 oc I (B) flies, oc' is a viable allele of otd caused by a chromosomal inversiondisruptingthe noncodingregionof the otd gene(Finkelsteinet al., 1990). In the oc mutant brain,a dorsomedialstructure,the protocerebral bridge, is missing (arrow).

Brain Segmentation and Head Segmentation In metameric patterning of the trunk regions of Drosophila, the activity of the segment polarity gene e n is essential for the establishment and maintenance of the posterior compartment of each segment (reviewed in Ingham and Martinez-Arias, 1992). Based on its function in trunk segmentation, e n expression has been used as a molecular marker to identify head segments in Drosophila (Cohen and J(Jrgens, 1991; Diederich et al., 1991; Schmidt-Ott

neuromere

bl I

en-bl

neuromere

en-b2 neuromere

en-b3 Figure 7. Summaryof the SegmentalOrganizationof the Embryonic Drosophila Brain Simplified schematicsummarydiagram showingthat the Drosophila embryonicbrain consistsof three neuromeres,bl, b2, and b3. Posterior bordersof each of these neuromeresare marked with stripes of en-expressingcells.The two partsof neuromerebl are interconnected by the preoralembryonicbraincommissure,which is formedjust anterior to the en-bl stripe. The two parts of neuromereb3 are interconnected by the tritocerebralcommissure,which is formed just anterior to the en-b3 stripe. The three brain neuromeresare connected by longitudinalaxontractsand connectivesthat extendfurther posteriorly to the subesophagealganglion.In otd mutant brains, neuromerebt, including the preoral brain commissure, is missing or severely deranged, whereas in ems mutant brains, neuromeresb2 and b3 are missing or severely deranged.

and Technau, 1992; J~rgens and Hartenstein, 1993; Schmidt-Ott et al., 1994). It has been suggested that the embryonic fly brain consists of four neuromeres (labral, ocular, antennal, and intercalary), based on an assignment of e n spots in the head region using conventional light microscopy (Schmidt-Ott and Technau, 1992). The problem of head segmentation in insects has been debated since the turn of the last century (reviewed in Rempel, 1975). The major obstacle for the resolution of this dispute is the lack of clear, morphologically visible landmarks. The complex process of head involution in Drosophila provides another obstacle to the analysis of the metameric organization of the cephalic region. In this report, we have focused directly on the metameric development of the brain, which is set aside from the process of head involution, and thus undergoes relatively simple developmental changes, allowing tracing of each neuromere in the course of cephalogenesis. Our data argue for the presence of three neuromeres in the embryonic brain for the following reasons. First, three distinct en-expressing stripes are observed in the embryonic brain; these cross each hemisphere at discrete levels along the neuraxis of the brain. Second, the three e n stripes delimit the sites of formation of three major axonal pathways: the preoral embryonic brain commissure, the antennal nerve root, and the tritocerebral commissure (for a review of the role of segmental borders in axon guidance cues, see Steindler, 1993). Third, the expression domains of the e m s and o t d homeobox genes largely respect the putative segmental borders set by the e n stripes, and mutational inactivation of these genes result in specific regional defects whose boundaries approximate the segmental borders defined by e n expression. These data do not exclude the possibility of further regionalization at the anterior end of the embryonic brain. Expression of o t d is not observed in the most anterior parts of the brain, and this region could correspond to the archicerebrum, which has been proposed previously based on classical morphological studies (Rempel, 1975). However, the identification of this anterior brain region

Neuron 776

as a distinct metameric structure awaits further genetic analysis. Functional Roles of the o t d and e m s Homeobox Genes in Brain Segmentation A great deal is known about the molecular genetic processes involved in segmentation of the main trunk region of the Drosophila embryo (reviewed in Akam, 1989). There, segmentation is controlled by a hierarchy of regulatory gene c a s c a d e s starting with the gradients of maternal gene products, which in turn are translated into broad embryonic domains by a set of gap genes. Subsequently, each of these initial embryonic domains are further subdivided by pair rule and segment polarity genes. Finally, segmental identity is specified by the homeotic genes of the A n t e n n a p e d i a and B i t h o r a x c o m p l e x e s . The roles of the o t d and eros homeobox genes in the segmentation of the cephalic regions, including the brain, appear to be different from the roles of other homeobox genes in trunk segmentation. Both o t d and eros are expressed at the early blastoderm stage as broad stripes under the control of the b i c o i d gradient, a feature characteristic of gap genes (Finkelstein and Perrimon, 1990; Diederich et al., 1991 ; Walldorf and Gehring, 1992). Moreover, in mutants o f o t d a n d eros, corresponding epidermal segmental structures (Cohen and JSrgens, 1990) and brain neuromeres (this work) are missing instead of being homeotically transformed toward other segmental identities. On the other hand, the persisting expression of the two genes up to late embryonic stages suggests that these genes might also act to specify the identity of each segment at later stages. These findings imply that the processes of metameric subdivision and segmental specification in the brain might be tightly coupled. Indeed, a possible dual function of o t d and e m s as gap-like and homeotic genes may explain, at least in part, the absence of pair rule gene expression in the head region including the developing brain. Although we have shown that e m s is required for the development of brain neuromeres b2 and b3, we still do not know which genetic processes confer individual identities to each of these two neuromeres. The labial gene, one of the most anteriorly expressed homeobox genes in the A n t e n n a p e d i a complex, is expressed in neuromere b3 and may be involved in its specification (Diederich et al., 1991; our unpublished data). Furthermore two other headspecific gap-like genes, b u t t o n h e a d a n d s l o p p y paired, are expressed at the early blastoderm stage in overlapping domains slightly posterior to those of the o t d and e m s genes (Cohen and JLirgens, 1990; Wimmer et al., 1993; Grossniklaus et al., 1994). Mutations of b u t t o n h e a d and s l o p p y p a i r e d result in loss of head epidermal structures, including the antennal and intercalary segments. Although neither of these genes contains a homeobox sequence, their early expression patterns suggest they may be involved in the metameric subdivision of the cephalic region including the developing brain.

Evolutionary Implications Recently, homologs of the o t d and eros genes, Otx1/2 and Emx1/2, have been cloned from mouse based on conservation of the homeobox sequences. These mammalian homeobox genes show regionalized expression patterns in the developing forebrain and midbrain as well as in the layers of the developing cerebral cortex and cerebellum (Simeone et al., 1992a, 1992b, 1993; Frantz et al., 1994). Moreover, both o t d and eros homologs have been isolated in Xenopus, zebrafish, and chicken (reviewed in Puelles and Rubenstein, 1993; Finkelstein and Boncinelli, 1994). The brain-specific expression patterns of these o t d / O t x and e m s l E m x homeobox genes in both flies and mammals suggest that the function of these genes in brain development might be evolutionarily conserved. If this is the case, targeted elimination of the Otx and E m x genes should lead to patterning defects in the mammalian brain. It has been proposed that the basic mechanisms of anterior patterning were established in a primitive ancestor of both flies and mammals, implying that cephalization evolved only once (Finkelstein and Boncinelli, 1994). In agreement with this hypothesis is the striking conservation in expression (and possibly in function) of the o t d l O t x and e m s l E m x homeobox genes in the development of the brain. Considering the evolutionary conservation that has already been shown for the H O M I l - l o x genes (see Kenyon, 1994; Keynes and Krumlauf, 1994), it seems likely that the basic mechanisms involved in building complex brains were established only o n c e in evolution. Experimental Procedures Fly Strains The wild-typestrain was Oregon-R.The followingmutant alleles and strains were used: ot~ A~°~,otdxcS~,otdx°87, otd YH13,In(1)L149 oc ~, oc ~a~, oc ~b(Wieschauset al., 1984; Finkelsteinand Perrimon,1990),ems 9"e~, ems '°A~z(Dalton et al., 1989), eros 7D99(Walldorf and Gehring, 1992), ems ~°°29(JQrgens et al., 1984), and fus4 P[en lacZ] (Kassis, 1990). Embryos were staged accordingto Campos-Ortegaand Hartenstein (1985) and Kl~mbt et al. (1991). Immunocytochemistry and Histology Embryos were dechorionated,fixed, and labeled according to Patel (1994) and Therianos et al. (1995). Primary antibodies were rabbit anti-HRP (FITC-conjugated)diluted 1:250 (Jan and Jan, 1982) (Jackson Immunoresearch),mouseanti-en diluted 1:1 (Patel et al., 1989), rat anti-emsdiluted 1:2000 (Walldorfand Gehring, 1992), rat anti-otd diluted 1:2 (Wieschaus et al., 1992), and rat anti-RK2 (repo) diluted 1:1250 (Campbell et al., 1994). Secondary antibodies were DTAFconjugated goat anti-rat, TRITC-conjugatedgoat anti-rat, and Cy3conjugatedgoat anti-mouse(Jackson Immunoresearch),all diluted 1: 250 (usuallyafter preabsorbtionagainstfixed embryos).Fluorescencelabeled whole-mountembryoswere mountedin Vectashield H-1000 (Vector). For sectioning,adult flies were fixed, dehydrated,imbedded in paraffin, orientedaccordingto neuraxis,and cut into 7 p.msections as describedby Heisenbergand B6hl (1979). Sectionswere mounted on coated glass slides, and neural structureswere visualized by autofluorescence. Light and Laser Scanning Confocal Microscopy For light microscopy, a Zeiss compound microscope (Axioskop) equippedwith fluorescenceand Nomarksioptics was used. Images were captured electronically using a MTI CCD-72Ex camera (Dage

Brain Development in Drosophila 777

MTI) and transferred on a Macintosh Quadra 800. Figures were arranged and labeled using Adobe Photoshop and printed using a Kodak XLS 8600 PS Printer. For laser confocal microscopy, a Leica TCS 4D was used. Optical sections ranged from 0.8 to 3 p.m and were recorded in the line average mode with picture size of 512 x 512 pixels. Images from optical sections were captured electronically, processed digitally, and then arranged, labeled, and printed as for light microscopy. Anatomical orientations of the brain are always according to neuraxie (see Therianos et al., 1995).

Acknowledgments We would like to thank R. Finkelstein, U. Walldorf, K. Basler, C. S. Goodman, and A. Tomlinson for kindly providing us with antibodies, fly strains, and cDNA clones; K. Ballmer-Hofer, J. Hagmann, and M. DOrrenberger for confocal microscopy facilities; and the Developmental Studies Hybridoma Bank (N ICHD, N01 -HD-2-3144). This work was supported by the Swiss National Science Foundation (to H. R. and K. F. T.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC Section t734 solely to indicate this fact. Received August 4, 1995.

References Akam, M. (1989). Hox and HOM: homologous gene clusters in insects and vertebrates. Cell 57, 347-349. Campbell, G., G(Sring, H., Lin, T., Spana, E., Andersson, S., Doe, C. Q., and Tomlinson, A. (1994). RK2, a glial-specific homeodomain protein required for embryonic nerve cord condensation and viability in Drosophila. Development 120, 2957-2966. Campos-Ortega, J. A., and Hartenstein, V. (1985). The Embryonic Development of Drosophila melanogaster (Heidelberg: Springer Verlag). Cohen, S. M., and JOrgens, G. (1990). Mediation of Drosophila head development by gap-like segmentation genes. Nature 346, 482-485. Cohen, S. M., and JOrgens, G. (1991). Drosophila headlines. Trends Genet. 7, 267-272. Dalton, D., Chadwick, R., and McGinnis, W. (1989). Expression and eml~ryonic function of empty spiracles: a Drosophila homeobox gene withtwo patterning functions on the anterior-posterior axis of the embryO. Genes Dev. 3, 1940-1956. Diederich, R. J., Pattatucci, A. M., and Kaufman, T. C. (1991). Development and evolutionary implications of labial, Deformed and engrailed expression in the Drosophila head. Development 113, 273-281. Finkelstein, R., and Boncinelli, E. (1994). From fly head to mammalian forebrain: the story of otd and Otx. Trends Genet. 10, 310-315. Finkelstein, R., and Perrimon, N. (1990). The orthodenticle gene is regulated by bicoid and torso and specifies Drosophila head development. Nature 346, 485-488. Finkelstein, R., Smouse, D., Capaci, T. M., Spradling, A. C., and Perrimon, N. {1990). The orthodenticle gene encodes a novel homeodomain protein involved in the development of the Drosophila nervous system and ocellar visual structures. Genes Dev. 4, 1516-1527. Frantz, G. D., Weimann, J. M., Levin, M. E., and McConnel, S. K. (1994). Otxl and Otx2 define layers and regions in developing cerebral cortex and cerebellum. J. Neurosci. 14, 5725-5740. Grossniklaus, U., Cadigan, K. M., and Gehring, W. J. (1994). Three maternal coordinate systems cooperate in the patterning of the Drosophila head. Development 120, 3155-3171. Guthrie, S. (1995). The status of the neural segment. Trends Neurosci. 18, 74-79. Halter, D. A., Urban, J., Rickert, C., Ner, S. S., Ito, K., Travers, A. A., and Technau, G. M. (1995). The homeobox gene repo is required for

the differentiation and maintenance of glial function in the embryonic nervous system of Drosophila melanogaster. Development 121,317322. Hartenstein, V. (1993). Atlas of Drosophila Development (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press). Heisenberg, M., and B6hl, K. (1979). Isolation of anatomical brain mutants of Drosophila by histological means. Z. Naturforsch. 34, 143147. Ingham, P. W., and Martinez-Arias, A. (1992). Boundaries and fields in early embryos. Cell 68, 221-235. Jan, L. Y., and Jan, Y. N. (1982). Antibodies to horseradish peroxidase as specific neuronal markers in Drosophila and grasshopper embryos. Proc. Natl, Acad. Sci. USA 79, 2700-2704. J0rgens, G, and Hartenstein, V. (1993). The terminal regions of the body pattern. In The Development of Drosophila, M. Bate and A. Martinez-Arias, eds. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press), pp. 687-746. J~3rgens, G., Wieschaus, E., Nesslein-Volhard, C., and Kluding, H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. I1. Zygotic loci on the third chromosome. Roux's Arch. Dev. Biol. 193, 283-295. Kassis, J. A. (1990). Spatial and temporal control elements of the Drosophila engrailed gene. Genes Dev. 4, 433-443. Kenyon, C. (1994). If birds can fly, why can't we? Homeotic genes and evolution. Cell 78, 175-180. Keynes, R., and Krumlauf, R. (1994). Hox genes and regionatization of the nervous system. Annu. Rev. Neurosci, 17, 109-132. Kl&mbt, C., Jacobs, J. R., and Goodman, C. S. (1991). The midline of the Drosophila central nervous system: a model for the genetic analysis of cell fate, cell migration, and growth cone guidance. Cell 64, 801-815. Lumsden, A. (1990). The cellular basis of segmentation in the developing hindbrain. Trends Neurosci. 8, 329-335. Patel, N. H. (1994). Imaging neuronal subsets and other cell types in whole mount Drosophila embryos and larvae using antibody probes. In Drosophila melanogaster: Practical Uses in Cell Biology, L. S. B. Goldstein and E. Fyrberg, eds. (New York: Academic Press). Patel, N. H., Martin-Blanco, E., Coleman, K. G., Poole, S. J., E~lis, M. C., Kornberg, T. B., and Goodman, C. S. (1989). Expression of engrailed proteins in arthropods, annelids, and chordates. Cell 58, 955-968. Puelles, L., and Rubenstein, J. L. R. (1993). Expression patterns of homeobox and other putative regulatory genes in the embryonic mouse forebrain suggest a neuromeric organization. Trends Neurosci. 16, 472-479. Rempel, J. G. (1975). The evolution of the insect head: the endless dispute. Quest. Entomol. 11, 7-25. Rubenstein, J. L. R., Martinez, S., Shimamura, K., and Puelles, L. (1994). The embryonic vertebrate forebrain: the prosomeric model. Science 266, 578-580. Schmidt-Ott, U., and Technau, G. M. (1992). Expression of en and wg in the embryonic head and brain of Drosophila indicates a refolded band of seven segment remnants. Development 116, 111-125. Schmidt-Ott, U., Gonzalez-Gaitan, M., Jeckle, H., and Technau, G. M. (1994). Number, identity, and sequence of the Drosophila head segments as revealed by neural elements and their deletion patterns in mutants. Proc. Natl. Acad. Sci. USA 91, 8363-8367. Simeone, A., Acampora, D., Gulisano, M., Stornaiuolo, A., and Boncinelli, E. (1992a). Nested expression domains of four homeobox genes in the developing rostral brain. Nature 358, 687-690. Simeone, A., Gulisano, M., Acampora, D., Stornaiuolo, A., Rambaldi, M., and Boncinelii, E. (1992b). Two vertebrate homeobox genes related to the Drosophila empty spiracles gene are expressed in the embryonic cerebral cortex. EMBO J. 11, 2541-2550.

Neuron 778

Simeone, A., Acampora, D., Mallamaci, A., Stornaiuolo, A., D'Apice, M. R., Nigro, V., and Boncinelli, E. (1993). A vertebrate gene related to orthodenticle contains a homeodomain of the bicoid class and demarcates anterior neuroectoderm in the gastrulating mouse embryo. EMBO J. 12, 2735-2747. Steindler, D. A. (1993). Glial boundaries in the developing nervous system. Annu. Rev. Neurosci 16, 445-470. Strauss, R., and Heisenberg, M. (1993). A higher control center of locomotor behavior in the Drosophila brain. J. Neurosci. 13, 18521861. Therianos, S., Leuzinger, S., Hirth, F., Goodman, C. S., and Reichert, H. (1995). Embryonic development of the Drosophila brain: formation of commissural and descending pathways. Development 121, 38493860. Truman, J. W., Taylor, B. J., and Awad, T. (1993). Formation of the adult nervous system. In The Development of Drosophila, M. Bate and A. Martinez-Arias, ed. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press), pp. 1245-1275. Walldorf, U., and Gehring, W. J. (1992). Empty spiracles, a gap gene containing a homeobox involved in Drosophila head development. EMBO J. 11, 2247-2259. Wieschaus, E., NL~sslein-Volhard, C., and JQrgens, G. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. 3. Zygotic loci on the X-chromosome and 4th chromosome. Roux's Arch. Dev. Biol. 193, 296-307. Wieschaus, E., Perrimon, N., and Finkelstein, R. (1992). orthodenticle activity is required for the development of medial structures in the larval and adult epidermis of Drosophila. Development 115, 801-811. Wimmer, E. A., J~ckle, H., Pfeifle, C., and Cohen, S. M. (1993). A Drosophila homologue of human Spl is a head-specific segmentation gene. Nature 366, 690-694. Xiong, W.-C., Okano, H., Patel, N. H., Blendy, J. A., and Montell, C. (1994). repo encodes a glial-specific homeodomain protein required in the Drosophila nervous system. Genes Dev. 8, 981-994.