The Interactive Fly
Genes involved in tissue and organ development
Determination of neuroblast identity in the neurectoderm
Control the temporal sequence of neuroblast specification
Temporal control of the development of neural sublineages
Using Fas2 to chart the structure of the neuropile
Separate sections of The Interactive Fly group genes according to their involvement in glia morphogenesis and axonogenesis.
The Drosophila central nervous system (CNS) develops from
a bilateral neuroectoderm that lies on either side of a narrow strip of
ventral midline cells. Single neuroectodermal cells
delaminate from the surface epithelium, in a fixed pattern, and move into the interior of
the embryo to form neural precursor cells called neuroblasts. The early
neuroblasts form an orthogonal grid of four rows (1, 3, 5, and 7) along
the anterior-posterior (AP) axis and three columns (ventral,
intermediate, and dorsal) along the dorsoventral (DV) axis.
Subsequently, each neuroblast expresses a characteristic combination of
genes and contributes a stereotyped family of neurons and glia to the
CNS. Thus the earliest steps in patterning the CNS are the formation
and specification of neuroblasts.
Neuroblast formation is regulated by two phenotypically opposite
classes of genes: Proneural genes promote neuroblast formation, whereas
the neurogenic genes inhibit neuroblast formation. Proneural genes
encode a family of basic helix-loop-helix transcription factors that
are expressed in 4-6 cell clusters at specific positions within the
neuroectoderm. Embryos lacking the proneural genes achaete/scute or lethal of scute have a
reduced number of neuroblasts (for review, see Skeath,
1994). Conversely, neurogenic genes are expressed uniformly in the
neuroectoderm, and embryos that lack any one neurogenic gene function,
such as Notch or Delta, develop an excess number of
neuroblasts (for review, see Campos-Ortega, 1995).
Neuroblast identity is determined in the neuroectoderm. Neuroblasts delaminate in five waves spanning approximately three hours. The generation of neuronal diversity begins with the specification of
unique neuroblast identities along both the anterior-posterior (AP) and dorsal-ventral (DV) axes. The pair-rule genes
wingless, hedgehog, gooseberry, and
engrailed are expressed in stripes of neuroectoderm that
subdivide the AP axis. These genes are required for establishing AP row
identity within the neuroectoderm and neuroblasts (Chu-LaGraff,
1993; Zhang, 1994; Skeath, 1995; Bhat, 1996; Matsuzaki, 1996; Bhat, 1997 and McDonald, 1997). For example, gooseberry is expressed in row 5 neuroectoderm. Embryos
lacking gooseberry function have a transformation of row 5 into row 3 neuroectoderm and neuroblast identity, whereas misexpression
of gooseberry results in the converse row 3 to row 5 transformation (Zhang, 1994 and Skeath, 1995). Similarly,
wingless encodes a protein secreted from row 5 and required
for specifying the fate of the adjacent rows 4 and 6 neuroectoderm and
neuroblasts (Chu-LaGraff, 1993). For information on the expression of segmentation genes and neuroblast identity genes in specific neuroblasts, see Chris Doe's Hyper-Neuroblast map.
Three genes are expressed in restricted domains along the DV axis within
the neuroectoderm: ventral nervous system defective (vnd) is an NK2 class homeobox gene expressed in the ventral
column neuroectoderm (Jimenez, 1995 and Mellerick,
1995) and muscle segment homeobox (msh) is a homeobox
gene expressed in the dorsal column neuroectoderm and neuroblasts
(D'Alessio, 1996 and Isshiki, 1997). Mutations in
vnd cause defects in neuroblast formation and lead to severe
defects later in neurogenesis (White, 1983 and Skeath, 1994).
Mutations in
msh result in a partial transformation of dorsal neuroblasts
into a more ventral or intermediate column identity, without affecting
neuroblast formation (Isshiki, 1997). Signaling via the EGF
receptor is required to establish ventral and/or
intermediate column fates in the neuroectoderm (Rutledge, 1992;
Raz, 1993; Schweitzer, 1995; Skeath, 1998; Udolph, 1998 and
Yagi, 1998). A newly cloned homeobox gene, intermediate neuroblasts defective (ind) is
the first gene known to be expressed specifically in the intermediate column of neuroectoderm and neuroblasts. ind
function is required for the establishment of intermediate column
identity in the neuroectoderm, and for the formation of intermediate
column neuroblasts. There is a hierarchical cascade of
transcriptional repression. Vnd represses ind
expression to establish the ventral boundary of ind
transcription, and ind represses msh to establish the
ventral boundary of msh transcription. The homeobox genes
expressed in columns within the Drosophila
neuroectoderm--vnd, ind, and msh--each have
gene homologs expressed in corresponding domains along the DV axis of
the vertebrate neural ectoderm. On this basis it appears that
fundamental molecular mechanisms of DV patterning may be similar in
Drosophila and vertebrates (Weiss, 1998 and McDonald, 1998).
Stage-specific inductive signals in the Drosophila neuroectoderm control the temporal sequence of neuroblast specification. To test when identity specification of the various neuroblasts takes place and whether extrinsic signals are involved, heterochronic transplantation experiments were performed. Single neuroectodermal cells from stage 10 donor embryos (after S2) were transplanted into the neuroectoderm of host embryos at stage 7 (before S1) and vice versa. The fate of these cells was examined by determining the makeup of their lineages at stage 16/17. Transplanted cells adjust their fate to the new temporal situation. Late neuroectodermal cells are able to take over the fate of early (S1/S2) neuroblasts. The early neuroectodermal cells preferentially generated late (S4/S5) neuroblasts, despite their reduced time of exposure to the neuroectoderm. Furthermore, neuroblast fates are independent from divisions of neuroectodermal progenitor cells. It is concluded from these experiments that neuroblast specification occurs sequentially under the control of non-cell-autonomous and stage-specific inductive signals that act in the neuroectoderm (Berger, 2001).
The segmented CNS (ventral nerve cord) of the Drosophila embryo is relatively simple, consisting of approximately 400 cells per hemineuromere. These originate after gastrulation from the ventral neurogenic region of the ectoderm. About 25% of the neuroectodermal cells delaminate into the embryo as CNS progenitor cells, called neuroblasts (NBs). The singling out of the NBs from among neuroectodermal cells is achieved by the activity of proneural and neurogenic genes. In each hemisegment approximately 30 NBs delaminate from the neuroectoderm according to a stereotyped spatiotemporal pattern. Each NB delaminates from a specific region of the neuroectoderm to occupy a particular place within the subectodermal NB layer. The process of delamination has been divided into five successive waves (S1-S5) with particular subpopulations of identified NBs delaminating during each wave. Thus, each NB is characterized by a typical position and time of delamination. Furthermore, each NB expresses a specific set of molecular markers. Finally, the unique identity of each NB is revealed by the production of a characteristic cell lineage (Berger, 2001 and references therein).
Crucial steps in the specification of the various NB identities appear to take place before delamination by the interpretation of positional information in the neuroectoderm encoded by segmentation genes and dorsoventral patterning genes. Heterotopic transplantation experiments have shown that neuroectodermal cells become committed by these spatial cues to different degrees. For example, whereas dorsal neuroectodermal cells are able to adjust their fate when transplanted to more ventral positions, ventral neuroectodermal cells exhibit firm commitment and produce lineages consistent with their origin. These experiments refer to a given developmental stage (early gastrula, stage 7). However, the time of delamination differs between NBs, and the identity of a given NB correlates with a certain time of delamination. This implies that NB specification requires temporal cues in addition to positional information (Berger, 2001 and references therein).
The mechanisms behind the temporal sequence of NB specification are unknown. Different modes of regulation could be envisaged. For example, all NB identities, including the respective times of delamination, might become firmly determined at an early stage and are cell-autonomously expressed during further development. Alternatively, progenitor cells might acquire NB-identities sequentially under the influence of extrinsic signals. To test whether the developmental potencies of neuroectodermal progenitor cells change over time and whether inductive signals are involved, the temporal axis was manipulated independently from spatial cues by performing heterochronic transplantations of neuroectodermal cells. Neuroectodermal cells were transplanted from stage 7 donors (early gastrula, before S1) into stage 10 hosts (after S2), and vice versa. The identities assumed by these cells were determined by analyzing their lineages in the host embryos at stage 16/17. In both experimental situations, neuroectodermal cells are able to adjust their fate to the new environment. Late neuroectodermal cells can generate early (S1, S2) NBs. Early neuroectodermal cells preferentially produced late (S3-S5) NB lineages, despite having been exposed to the neuroectoderm for a significantly reduced period of time. Late NB fates are independent of previous divisions of neuroectodermal progenitor cells. These data suggest that extrinsic inductive signals exist in the neuroectoderm that change over time to control the specification of temporal subsets of neuroblasts (Berger, 2001).
In one set of experiments, neuroectodermal cells from stage 10 embryos were heterochronically transplanted into the neuroectoderm of 2 hours younger, early gastrula (stage 7) hosts. The transplanted cells gave rise to CNS clones, or to epidermal clones, or to mixed CNS/epidermal clones. This shows that despite their more advanced age, the implanted cells participate in the cell interaction process that leads to the decision of neurectodermal cells between an epidermogenic and a neurogenic fate. Remarkably, however, among the cells that follow the neural pathway, about 50% produced lineages typical for early NBs (S1, S2), as for example, NB1-1, MP2, NB2-2 or NB4-2. This indicates that neuroectodermal cells at stage 10, which normally only give rise to late NB lineages, have not lost the potency to assume identities of early NBs. Taken together these data indicate that late ectodermal cells (stage 10) are not irreversibly specified, and that signals exist in the early neuroectoderm (stage7) that are sufficient to induce early NB fates. Thus, instead of being merely based on cell-autonomous properties, the temporal regulation of early NB determination appears to be mediated by extrinsic inductive signals that are active in the early neuroectoderm (Berger, 2001).
Reduced time of exposure to the neuroectoderm does not prevent formation of late NBs. Having shown that the determination of early NB fates depends on stage specific inductive signals, whether inductive signals are also involved in the generation of late NB fates was tested. Cells from the early neuroectoderm (stage 7) were heterochronically transplanted into the neuroectoderm of stage 10 host embryos. Among 132 identifiable clones obtained from these cells, 24 (19%) were CNS clones and 108 (81%) epidermal clones. Closer analysis of the 24 CNS clones revealed that about 80% (n=19) of them corresponded to lineages typical for late NBs, like 2-1, 5-4, 6-4 or 7-3, and only 20% (n=5) to early NB lineages. Therefore, the transplanted cells tend to adopt to the new temporal environment regarding the identities of NBs to be formed. Although having skipped two hours of exposure to the neuroectoderm, a significant proportion of them can compensate for this lack of time. Thus, the cells are not bound to an intrinsic timer to become specified as late NBs, but are able to react to inductive signals in the late neuroectoderm. The 20% of cells that developed an early NB fate might point to differences in the degrees of commitment of neuroectodermal cells at a given stage or to an insufficient exposure to signaling in the late neuroectoderm under the experimental conditions (Berger, 2001).
Determination of late NBs does not depend on previous division in the neuroectoderm. As opposed to early NBs the lineages of S4 and S5 NBs, and some of the S3 NBs have an epidermal sister clone. This is due to the postblastodermal division pattern of neuroectodermal progenitors. Progenitors developing as S1 and S2 NBs do not divide before delamination from the neuroectoderm: some of those giving rise to S3 NBs divide, and those giving rise to S4 and S5 NBs (NBs 1-3, 2-1, 2-4, 3-3, 4-3, 4-4, 5-1, 5-4, 5-5 and 7-3) always divide in the neuroectoderm. Only one of the daughter cells that results from this division subsequently delaminates as a late NB, the other remains in the periphery to develop as an epidermoblast. Is this neuroectodermal division required for late NBs to form and become properly specified? When neuroectodermal cells are heterochronously transplanted from stage 7 donors into stage 10 hosts, they are deprived from the phase in which the first wave of divisions normally runs through the neuroectoderm. Most of the CNS clones obtained from these cells corresponded to lineages of late NBs. However, whereas S4 and S5 NBs normally have an obligatory epidermal sister clone, the situation is variable under the experimental conditions. Some of these clones have a sister clone consisting of epidermal cells, whereas the other clones lack an epidermal sister clone (Berger, 2001).
These data show that: (1) proliferation of individual neuroectodermal progenitors can be influenced by surrounding tissue; (2) late NBs can segregate from the neuroectoderm without having previously divided; and (3) late NBs do not depend on a previous division to acquire an individual identity and to produce their specific and complete CNS lineage. These observations lend further support to the idea that the temporal pattern of NB determination depends on inductive signals in the neuroectoderm instead of following a stereotype cell autonomous clock (Berger, 2001).
There is ample evidence that the specification of NBs crucially depends on positional information in the neuroectoderm provided by the products of segmentation genes and dorsoventral patterning genes. Part of this information becomes integrated into the cell-autonomous program of the cells before neurogenesis. Another part, however, is subsequently provided by extrinsic signals. For example, the segment polarity gene wingless (wg) is segmentally expressed in a single row of neuroectodermal cells and the secreted Wg protein is non-autonomously required in adjacent anterior and posterior neuroectodermal cells for the formation and specification of NBs. Along the dorsoventral axis, the secreted Spitz and Vein proteins are involved in conferring NB identities. These heterochronic transplantation experiments show that extrinsic signals are also involved in NB specification along the temporal axis. Although neuroectodermal cells of stage 10 embryos normally never produce NBs belonging to the group of S1 and S2 NBs, they do so after being transplanted into stage 7 neuroectoderm. The possibility that the cells follow this fate autonomously after being released from signals that normally inhibit these fates in the late neuroectoderm is incompatible with the following evidence. Cells from the non-neurogenic dorsal ectoderm of stage 10 donors are able to adopt a CNS fate upon heterotopic transplantation, and to become specified as early NBs. However, they are unable to autonomously develop as a NB in cell culture. Thus, the transplanted late cells do react to signals in the early neuroectoderm and adjust their development accordingly. This also seems to be possible in the other direction. Upon transplantation of stage 7 neuroectodermal cells into the neuroectoderm of hosts at stage 10, most of the CNS lineages obtained are typical for NBs that normally delaminate late. Similar to the situation in Drosophila, heterochronic transplantations using the developing ferret brain have revealed an interaction scenario of extrinsic cues and intrinsically changing properties for the sequential birth of neuronal cell types from ventricular zone progenitor cells. Progenitor cells from very young embryos can adjust their fate to older host tissues. By contrast, cells from older tissue transplanted into younger host brains adopt only fates typical of their origin. The latter experiment reveals an irreversible intrinsic change of the developmental properties of older cells. Intrinsic changes over time are likely to occur also in the Drosophila neuroectodermal cells; however, they are reversible under the influence of external signals. It remains to be tested as to how far this is also the case for NBs once they have delaminated from the neuroectoderm (Berger, 2001 and references therein).
These experiments suggest that the entire temporal sequence of delamination of specific subsets of NBs is not readily determined in the early neuroectoderm but is controlled by the dynamic expression of stage specific signals. Segment-polarity genes play an important role in the formation and identity specification of NBs. They are segmentally expressed in particular rows of neuroectodermal cells. As the expression domains of some of these genes evolve dynamically and, hence, differ at different stages, they are also good candidates for being involved in the temporal control of NB formation/specification. The differential commitment of the late neuroblasts NB 6-4 (S3) and NB 7-3 (S5) has been shown to be mainly controlled by the interplay of the segment polarity genes naked (nkd) and gooseberry (gsb). Mutation of either nkd or gsb leads to the transformation of one NB fate to the other. Interestingly, however, the temporal sequence of their delamination is maintained, i.e. independent from these genes. This suggests that formation and specification of these two NBs is under independent control. Further work will have to test whether this is also the case for other NBs and to uncover the signals that regulate the temporal pattern of NB fate determination (Berger, 2001 and references therein).
What mechanisms control the sequential generation of neurons -- that is, how are unique fates acquired by the successive daughters of a neuroblast? After a neuroblast has been specified positionally, by the actions of segment polarity genes and a network of homeodomain genes, it delaminates from the
ectoderm and begins to divide unequally into one large and one small daughter cell. The large cell (still called a
neuroblast) continues to go on this way for a variable number of rounds. The small cell, called a ganglion mother
cell (GMC), typically divides equally one more time to generate a pair of postmitotic neurons. Often these
neurons form a stack on top of the neuroblast from which they originated. As postmitotic neurons in the insect
CNS do not generally migrate, the position of a neuron in the CNS depends on whether it was generated early
or late. In this way, a histogenetic order is built into the cellular cortex of the insect CNS, with early neurons
deep and close to the neuropil and late neurons next to the surface of the brain (Kambadur, 1998; Harris, 2001).
This arrangement of cells according to relative birth date is also observed in laminated structures in the vertebrate CNS, the best example being the cerebral cortex. In the mammalian cortex, cells acquire their fates at
the ventricular surface at the time they are born, and these postmitotic neurons cells then migrate to their
specified laminar destinations (McConnell, 1995). In both the cortex and the retina, it is thought that
progenitors are pluripotent and realize their particular fates by being exposed to an extracellular environment that
changes with time (Harris, 1997). Whether intrinsically or extrinsically controlled, particular combinations of
transcription factors are expressed in the neuroblasts of both the vertebrate retina and fly CNS over the course
of development, and these factors appear to restrict the competence of neuroblasts to the fates that are
appropriate (Harris, 2001 and references therein).
The first insights into this problem in the fly CNS were made in Odenwald's laboratory at the NIH. They
showed that expression of the transcription factor genes hunchback (hb), pdm, and castor (cas) occur
sequentially in the embryonic CNS of Drosophila (Kambadur, 1998). Furthermore, the CNS
neuroblasts themselves sequentially express these three genes in a conserved order (Brody,
2000), and whichever of the three genes is expressed in the neuroblast when it divides continues to be
expressed in the progeny. The transcription factor Grainyhead (Gh) appears to mark the NB after it has generated lineages marked by Hb, Pdm and Cas, and the Gh positive NB also generates a fourth lineage (Brody, 2000). Thus, the earliest generated neurons in the fly CNS tend to express hb, while later
generated neurons express pdm, and still later generated neurons express cas followed by gh. By analogy to the spatial
coordinate genes, these can be called 'temporal coordinate genes (Harris, 2001).
Isshiki (2001), working in Doe's laboratory, followed individual neuroblasts and their progeny. It appears that each neuroblast examined express four temporal coordinate genes -- hb, Kruppel (Kr), pdm, and cas, in that invariant order. By following the GMCs and their daughter neurons, Isshiki confirmed on a cellular level that each GMC maintains the expression profile of the temporal coordinate genes that its parent neuroblast displayed at the time the GMC was generated. The relevance of these temporal coordinates to neuronal fate was addressed with misexpression constructs and loss of function mutants in hb and Kr. These experiments led to respecification of GMCs and their progeny to earlier or later fates, as expected if these genes really are important to fate. Thus, these temporal coordinate genes play a similar role in fate specification along a histogenetic axis as the spatial coordinate genes play in the positional axes (Harris, 2001).
What, one might wonder, are the mechanisms that could generate these temporal transitions in the parent neuroblast? The first possibility is that global temporal cues, such as circulating hormones, or intracellular signaling such as Notch, Egf, Wingless, or Hedgehog, trigger these transitions. Since transitions in transcription factor expression take place in isolation, in cultured clones, it is concluded that once NBs initiate lineage development no additional signaling between NBs and the neuroectoderm and/or mesoderm is required to trigger the temporal progression of transcription factor expression during NB outgrowth (Brody, 2000; Harris, 2001).
Second, the cascade mechanism itself, in which each of these genes is responsible for turning on the next in the series, functions to regulate the transcription factor transitions. Overexpression of Hb activates Kr
and represses Pdm and Cas; overexpression of Kr activates Pdm, represses Cas, but has no effect on Hb expression; Pdm positively regulates Cas expression; and Cas repressed Pdm expression (Kambdadur, 1998 and Isshiki, 2001). Although there are appropriate sequential regulatory interactions of this kind, mutations in any of the earlier genes only subtly affect the temporal expression of subsequently expressed genes. Thus, although these interactions refine sequential expression, there must be additional elements to temporal regulation. Isshiki (2001) has shown that individual NBs go through the same sequence of expression in their own sweet time, independent of the developmental stages at which they delaminate. A third possibility is therefore suggested, a more mysterious clock mechanism may also be responsible for generating the order. The clock in this case appears to be directly related to the cell cycle, since arresting cell division with the Cdc25 mutant, string, freezes the pattern in time (Cui, 1995; Weigmann, 1995; Harris, 2001 and references therein).
The addition of a time axis adds to our understanding of how different neuronal types arise in the Drosophila CNS, and it also raises the intriguing problem of how multiple inputs regulate the expression of the temporal coordinate genes (Harris, 2001).
Patterns of embryonic neurogenesis in a primitive wingless insect, the silverfish; comparision with those in flying insects Neurogenesis was examined in the central nervous system of embryos of the primitively wingless
insect, the silverfish, Ctenolepisma longicaudata, using staining with toluidine blue and the
incorporation of bromodeoxyuridine. The silverfish has the same number and positioning of
neuroblasts as seen in more advanced insects and the relative order in which the different neuroblasts
segregate from the neuroectoderm is highly conserved between Ctenolepisma and the grasshopper,
Schistocerca. Of the 31 different neuroblasts found in a thoracic segment, one (NB 6-3) has a much
longer proliferative period in silverfish. Of the remainder, 14 have similar proliferative phases, while 16
neuroblasts have extended their proliferative period by 10% of the duration of embryogenesis (10%E) or greater in the
grasshopper, as compared with the silverfish. Both insects have similar periods of abdominal
neurogenesis except that in the silverfish terminal ganglion, a prominent set of neuroblasts continues
dividing until close to hatching, possibly reflecting the importance of cercal sensory input in this insect.
This comparison between silverfish and grasshopper shows that the shift from wingless to flying
insects was not accompanied by the addition of any new neuronal lineages in the thorax. Instead,
selected lineages undergo a proliferative expansion to supply the additional neurons presumably
needed for flight. The expansion of specific thoracic lineages was accompanied by the reduction of the
terminal abdominal lineages, specifically NB6-3, as flying insects began to de-emphasize their cercal sensory system. This neuroblast is notable in that it is the only neuroblast from the grasshopper set that is missing in Drosophila. A reasonable speculation is that NB 6-3 makes interneurons that deal with ascending information from the cercus (Truman, 1998).
Although abdominal neurogenesis in Ctenolepisma is roughly equivalent to that in the grasshopper, except in the terminal ganglion where it is higher, in the thorax it falls well below that seen in the grasshopper. In the silverfish, the last thoracic NBs stop dividing by about 70% of the way through embryonic development (70%E), whereas in grasshopper embryos, selected thoracic NBs continue dividing until slightly after 90%E. Comparison of the neurogenic periods between individual grasshopper and silverfish neuroblasts shows that only some of the neuroblasts have participated in this expansion. The greatest differences are seen for NBs 1-1, 5-1, and 6-4; these proliferate for 25%E to more than 30%E longer in the grasshopper as compared to the silverfish. An additional 14 neuroblasts have their proliferative period extended by at least 10%E in grasshopper embryos. The one neuroblast that goes counter to the overall trend is NB6-3, which has a proliferative period that extends for over 20%E longer in the silverfish. Hence, this lineage has become smaller in more advanced insects rather than becoming larger or staying the same. It is argued, however, that an increase in lineage size alone is not sufficient to conclude that that particular lineage produces neurons associated with flight (Truman, 1998).
Insect neurons are individually identifiable and have been used successfully to study principles of the formation and function of neuronal circuits. In Drosophila, studies on identifiable neurons can be combined with efficient genetic approaches. However, to capitalize on this potential for studies of circuit formation in the CNS of Drosophila embryos or larvae, it is necessary to identify pre- and postsynaptic elements of such circuits and describe the neuropilar territories they occupy. A strategy for neurite mapping is presented, using a set of evenly distributed landmarks labelled by commercially available anti-Fasciclin2 antibodies that remain comparatively constant between specimens and over developmental time. By applying this procedure to neurites labelled by three Gal4 lines, neuritic territories are shown to be established in the embryo and maintained throughout larval life, although the complexity of neuritic arborizations increases during this period. Using additional immunostainings or dye fills, Gal4-targeted neurites can be targetted to individual neurons and they can be characterized further as a reference for future experiments on circuit formation. Using the Fasciclin2-based mapping procedure as a standard (e.g., in a common database) would facilitate studies on the functional architecture of the neuropile and the identification of candiate circuit elements (Landgraf, 2003).
Working with defined pre- and post-synaptic neurons is a prerequisite for the study of mechanisms that underlie circuit formation. The fact that such neurons establish synaptic contacts with one another requires that some of their neurites project to a common region. Thus, proximity of neurites is a criterion that can be used towards the identification of putative pre- and postsynaptic neurons. In Drosophila (like in other insects), synaptic contacts are restricted to the neuropile, a cell body-free area, which also contains the ascending, descending, and commissural fibers. Unlike the gray matter in the vertebrate spinal cord (where cell bodies and synapses are intermingled), neuronal cell bodies of the Drosophila CNS are restricted to the synapse-free 'cortex' from where they send monopolar projections into the neuropile. These neuropilar accumulations of neurites of CNS neurons (i.e., efferent and interneurons) are joined by projections from peripheral sensory neurons. The functionality of thus established neuronal circuits demands that the spatial arrangements of synapse-bearing neurites in the neuropile are fairly reproducible between different individuals, as has been learned from analyses in larger insects. In order to map these reproducible neurites in the Drosophila neuropile, predominantly anatomical landmarks of the neuropile have been used to date as reference points for the relative positions of neuronal projections. Such landmarks are segmentally repeated nerve roots and commissures, or easily identifiable fiber tracts (so far applied only in the imaginal CNS (Landgraf, 2003).
This study capitalizes on a set of axon tracts that are labelled by the commercially available antibodies against the intracellular domain of Fasciclin2. These provide a set of standard landmarks that are evenly distributed throughout the neuropile. As shown by double-labellings with presynaptic markers, all Fasciclin2-positive fiber tracts are fully contained within the synaptic neuropile. They can be used in a very easy and efficient way for the charting of neurites in the neuropile. So far, Fasciclin2 fibre tracts have served as one-dimensional (mediolateral) landmarks in younger embryos. This approach has been extended by using the set of Fasciclin2 tracts in three dimensions and at different developmental stages. These analyses were exclusively centered on abdominal neuromeres for two reasons: predominantly, the abdominal motorsystem contributes to larval movement, and abdominal neuromeres face only minor reorganization during larval life (Landgraf, 2003).
Each Fasciclin2 fascicle has been named according to its relative dorsoventral (D, dorsal; C, central; V, ventral) and mediolateral (M, medial; I, intermediate; L, lateral) position. Such a nomenclature is neutral and can therefore be applied to any set of axon fascicles. The pattern of Fasciclin2 tracts remains relatively constant throughout larval development and thereby permits comparisons and extrapolations across different developmental stages. The main change to the embryonic pattern of Fasciclin2 in the neuropile is the addition of further elements, particularly five transverse projections (TP1-5) per neuromere in larval stages, which provide added reference points for the anteroposterior axis. From their association with different motor axons in the larva (TP1 with RP2 and VUM; TP2 with aCC), it is concluded that TP1 represents the pISN and TP2 the aISN nerve root (Landgraf, 2003).
In order to facilitate comparisons with published work, attempts were made to relate the Fasciclin2 pattern of the late embryo and larval stages to existing descriptions. For example, the Fasciclin2 pattern has frequently been used for work on the ventral nerve cord of earlier embryos, usually at 13 h of development. At this stage, three tracts can be resolved in the horizontal plane, of which the intermediate Fasciclin2 tract is formed or at least joined by axons of the MP1-interneurons (targeted by C544-Gal4), the medial tract by MP2-interneurons (targeted by 15J2-Gal4;. A split of the three tracts into vertically distinguishable bundles occurs during the next ca. 3 h. During this interval, it is still possible to trace the MP2/pCC- and MP1-axons via the C544- and 15J2-Gal4 lines when visualizing the Gal4-expressing neurons with the Uas-CD8-GFP reporter gene; later their Gal4-expression patterns change dramatically. Thus, despite the highly dynamic changes in the neuropile during this period (i.e., nerve cord condensation, closer apposition of neuropile at the midline and the fact that the intracellular Fasciclin2 domain vanishes from many cell bodies and axons), it is possible to map the MP2/pCC-axons to the DM (dorsointermediate) axons, and the MP1 interneuron axons to the dorsal CI-fascicles (Landgraf, 2003).
Classical neurobiological work on neuronal circuitry in other insects has been based on mapping strategies that used morphologically distinguishable axonal tracts in the neuropile and relates these to projection patterns of neurons. Similar strategies have been used for the thoracic adult CNS of Drosophila. The DM- and VMd-fascicles serve as reliable landmarks for distinguishing dorsal, intermediate, and ventral commissural tracts. The distinct patterns of sensory projections of different modalities, linked to the classical neurobiological literature, reveal a partitioning of the larval Drosophila neuropile. In an effort to relate the pattern of Fasciclin2 tracts to neuropilar regions, use has been made of three different Gal4 lines that target different subpopulations of sensory neurons (C161-, MJ94-, MzCh-Gal4). Sensory projections are confined to ventral regions, while neurites of motorneurons occupy the very dorsal neuropile. Thus, there is little, if any, physical overlap and contact between afferent sensory projections and central motorneuron neurites during larval stages. However, some overlap might occur lateral to the DM-fascicle, most likely with projections of the dbd and vbd-neurons. Thus, the data suggest that there are few, if any, monosynaptic connections between sensory and motorneurons in the embryonic and abdominal larval ventral nerve cord of Drosophila. However, this is a fairly rough estimation that will have to be tested by more detailed studies in the future (Landgraf, 2003).
Having described some spatial features of the neuropile with the help of the Fasciclin2 pattern, this charting strategy was next applied to three selected Gal4 driver lines. This effort is intended to identify and characterize neurons that are genetically amenable and that could be used for the investigation of neural circuit formation in the embryonic and larval Drosophila CNS. Three neural Gal4 lines were analyzed with precision. Before presenting detailed characteristics of abdominal Gal4-labelled neurons, an overview of the three Gal4 lines is provided: Per abdominal half-neuromere eve-Gal4RRK expresses Gal4 in two motor-(aCC and RP2) and one interneuron (pCC). DDC-Gal4 displays 9-11 Gal4-neurons, and MzVum-Gal4 12-14 cells plus 3 efferent VUM (Ventral Unpaired Median) neurons located at the ventral midline. In all three lines, Gal4 expression occurs in a defined sequence, and for most cells it is yet unclear to what extent a late onset of expression reflects a late birth and/or differentiation of those cells. Only in the aCC, pCC and VUM neurons is Gal4 expression initiated at the time of their respective births, thus making them amenable to genetic manipulations of axonal pathfinding and differentiation. Next, the relative strengths of Gal4 expression were compared and overall MzVum-Gal4 expresses strongest, followed by DDC-Gal4 and eve-Gal4RRK. However, Gal4 levels of different neuronal subsets in each Gal4 strain can differ significantly (e.g., in MzVum-Gal4, GABAergic interneurons express low levels while VUM and leucokinin-1-positive neurons express high levels). Because of differences in timing and strength of expression, Gal4-based manipulations would not be expected to affect all cells alike (Landgraf, 2003).
By virtue of the Fasciclin2-positive landmarks, it was possible to work out detailed descriptions of the neuropilar positions of neurites labelled by the three Gal4 lines eve-Gal4RRK, MzVum-Gal4, and DDC-Gal4. These studies clearly show that the Fasciclin2 framework allows spatial relationships between neurites to be pinpointed even across specimens: for example, neurites of the aCC and RP2 neurons (eve-Gal4RRK) are concentrated to form an oval in each hemineuromere that is located at the level of the DL-fascicle, medial to the ascending section of transverse projection 2 and anterior to transverse projection 1. At the same level (of the DL-fascicle), MzVum-Gal4-labelled neurites form whirlwind-like arrangements that have oval holes in their centers. These holes map to the region where aCC and RP2 neurites are concentrated, as indicated by the transverse projection 2. Thus, by using Fasciclin2-positive tracts as landmarks, spatial relationships of neurites are reproducibly revealed in three dimensions (Landgraf, 2003).
Since the pattern of Fasciclin2-positive axon tracts remains relatively constant from the late embryo to larva, it can also be used to investigate how neuronal projections change during this developmental period. The larval patterns of neurites described above are prefigured in the late embryo. For example, at late embryonic stages, the central arborisations of aCC and RP2 at the level of the DL-fascicle are also concentrated anterior to pISN (equals TP1 in the late larva), which corresponds to the region that lacks neurites in MzVum-Gal4 embryos. Thus, the principle spatial relationships between these sets of neurites (of aCC and RP2 versus those of MzVum-Gal4) appear to be laid down in the late embryo and maintained to larval stages, though the complexity and the spread of neurites increases over developmental time. This is an important observation because it suggests that: (1) By late embryonic stages neuritic arbors define those territories in the neuropile from which they will elaborate and spread during subsequent larval stages. Thus, principle spatial relationships between neurites are laid down during embryogenesis. (2) Data on the distribution of neurites obtained at one stage of development can be extrapolated and used to interpret other stages (Landgraf, 2003).
As demonstrated so far, using Fasciclin2 stainings significantly improves descriptions of the characteristic patterns of central neurites targeted by different Gal4-lines. However, these patterns of neurites are composites of different neurons. With this in mind, attempts were made to define methods with which to resolve such complex neuritic patterns into their constituent parts (Landgraf, 2003).
The first approach to tackle this problem is to employ antisera, which would reveal the morphologies of particular subsets of neurons. By using antibodies against the neurotransmitter/neuromodulator Serotonin and the neuropeptides Corazonin and Leucokinin-1 on nerve cords displaying Gal4-driven CD8-GFP expression, it is possible to define these neurites among the composite of Gal4-targeted projections that correspond to Serotonin, Corazonin, and Leucokinin-1 immunoreactive neurons and the regions of the neuropile that these occupy. Anti-Serotonin stains two and anti-Corazonin one neuron per hemineuromere. These three cells are targeted by the DDC-Gal4 line and appear to give rise to most of the DDC-Gal4-labelled neurites in the abdomen. Anti-Leucokinin-1 labels Gal4-targeted efferent projections forming type-3 terminals on the VL1-muscle (type-3v, DDC-Gal4;) and on the segment border muscle (type-3u, MzVum-Gal4). Of these, only the type-3u neuron is revealed by Leucokinin-1-like immunoreactivity and can thereby be traced back to a ventrolateral cell body in the CNS extending side branches toward the VL-fascicle. There are additional Leucokinin-1-positive projections associated with the DM-fascicle that are not targeted by MzVum-Gal4 but seem to originate from 2-4 (Gal4-negative) neurons at the anterior tip of the nerve cord. This has been confirmed by targeting the cytotoxin Ricin to MzVum::CD8-GFP neurons. This selectively abolishes all MzVum-Gal4-specific CD8-staining and the VL- but not the DM-associated Leucokinin-1-like immunoreactivity (Landgraf, 2003).
In summary, it has been shown that a small range of antisera can readily be used to reveal the projections of particular subsets of neurons. Such specific stainings are well suited to serve as spatial reference points in their own right. Moreover, in this instance, the Serotonin, Corazonin, and Leucokinin-1 immunoreactive neurons were instrumental in revealing some of the constituent parts of the complex projection patterns of the DDC-Gal- and MzVum-Gal4-lines (Landgraf, 2003).
Next, efferent neurons targeted by the three Gal4 lines were characterized and their axonal projections (nerve root and branch), target muscles, and terminal types were described. Based on morphological, molecular, and ultrastructural characteristics of motor terminals, several types of efferent neurons can be distinguished in Drosophila. It should be emphasised that distinctions between terminal types are not only of importance to studies of the Drosophila motor system but also correlate with differences between the central dendritic arbors of particular efferent neuron types. To classify the Gal4-labelled efferent neurons with respect to terminal type, a range of immunohistochemical stainings was employed: Synaptotagmin, Cysteine string protein, and Synapsin all represent presynaptic proteins involved in regulation of synaptic vesicles; Discs large is a predominantly postsynaptic protein, which labels the subsynaptic reticulum; and anti-Leucokinin-1 antisera detect an insect neuropeptide (Landgraf, 2003).
While the visualization of Gal4-labelled neurites via immunostaining is efficient, it is at the same time limited to particular subsets of cells, leaving many neurons unidentified. This limitation can be overcome by using standard neuronal tracers. To reveal the morphologies of those Gal4::CD8-GFP neurons, the neural tracer dye Cascade Blue was iontophoretically applied to individual cells. Thus, it was possible to define the positions of somata and central projections of all efferent neurons and a number of interneurons (Landgraf, 2003).
For the efferent neurons, it was found that their central projections are restricted to the dorsal neuropile (dorsal to the CI-fascicles). The only exception to this was the efferent SBM-neuron (MzVum-Gal4; whose short central arbors reside in the ventral neuropile where they associate with the VL-fascicle (consistent with Leucokinin-1 staining). In addition, it was found that differences in terminal type are reflected by distinctions in the central arbors of efferent neurons: while type-1 motoneurons elaborate extensive dendritic arbors (aCC and RP2; VA), efferent neurons with type-2 and type-3 terminals form comparatively sparse and stunted central arbors (VUM and SBM; VL1). Finally, these analyses suggest that the central projections of the same motoneuron in consecutive neuromeres do not overlap, i.e., they seem to behave in accordance with the tiling principle (Landgraf, 2003).
The interneurons of two of the Gal4-lines have been identified previously: pCC (eve-Gal4RRK) lies adjacent to the aCC motorneuron; three interneurons of DDC-Gal4 are serotonergic or corazonergic. In addition, two MzVum-Gal4 interneurons were identified via Cascade blue fill. These two intersegmental interneurons seem to contribute to most or all MzVum-Gal4-targeted neurites in the ventral neuropile, ventral to the CI-fascicle (except for intersegmental ascending and descending projections and the leucokinin-1-positive neurites associated with the VL-fascicle). It is possible that additional ventral neurites might be derived from the mVg- and GABAergic neurons of MzVum-Gal4 (Landgraf, 2003).
In summary, it was found that neurites targeted by MzVum-Gal4 segregate into a dorsal fraction, consisting primarily of motoneuronal side branches, and a ventral fraction derived almost exclusively from interneurons. This pattern simplifies interpretations of experimental results obtained with this Gal4-line (for example, if mutant backgrounds reveal selective impairment of only dorsal or ventral neurites). Having applied a combination of a standardized set of Fasicilin2-positive landmarks, specific antisera, and single cell tracings, it has been possible to (1) assign most neurites of the Gal4-lines to identified neurons, and (2) define the regions of the neuropile that they occupy. Future applications of a standardised mapping strategy to other Gal4 lines will considerably advance the understanding of the functional architecture of the Drosophila neuropile, and it will form a basis with which candidate pre- and postsynaptic circuit elements can be identified (Landgraf, 2003).
An important aspect of this study is that despite its limited scope it reveals an apparent partitioning of the neuropile into (possibly functionally) distinct regions. Facets of a functional architecture of the neuropile have already been documented such as the modality-specific sensory projections that partition the ventral neuropile. Due to the Fasciclin2-based mapping, these areas can now be named and the regions can be related to projection patterns of other neurons. In accordance with work published for other insects, the dorsal neuropile is predominantly occupied by the central arbors of efferent neurons (with the single exception of the ventral type-3u neuron arbors). There is little overlap with sensory areas so that direct connections between sensory and motor neurons will be the exception. In addition, different efferent neurons elaborate their central arbors in distinct anteroposterior regions of the dorsal neuropile. These territories seem to be defined in the embryo and they are maintained through larval stages, although areas of overlap between formally distinct territories increase as central arbors become more elaborate over time. This relative constancy of the topography of the neuropile over time also exists for Serotonin-, Corazonin-, and Leucokinin-1-positive neurons. An important consequence of such constancy for future research work is that neurites can be compared or descriptions extrapolated across different developmental stages (Landgraf, 2003).
Interestingly, neuropeptidergic projections seem to cluster in particular areas. Corazonin and Leucokinin-1 (and also Serotonin) are closely associated with the DM-fascicle. Published data suggest that antibodies against FMRF, molluscan neuropeptide SCPB, and Substance-P reveal neural structures that might also be localized in this median area. A second neuropeptide 'hot spot' is the VL-fascicles, where staining with anti-Leucokinin-1 antibodies is found. Also antisera against Allatostatin and Insulin appear to stain in this region. The fascicles are innervated by the posterior ascending cells of MzVum-Gal4 and DDC-Gal4 and curiously are detected with antisera against muscle myosin heavy chain. Interestingly, both of these neuropeptidergic 'hot spot' areas bear very prominent Fasciclin2-labelled neurites (Landgraf, 2003).
Berger, C., Urban, J. and Technau, G. M. (2001). Stage-specific inductive signals in the Drosophila neuroectoderm control the temporal sequence of neuroblast specification. Development 128: 3243-3251. Medline abstract: 11546741
Bhat, K.M. (1996). The patched signaling pathway mediates repression of gooseberry allowing
neuroblast specification by wingless during Drosophila neurogenesis. Development 122:
2921-2932. Medline abstract: 96379752
Bhat, K.M. and P. Schedl. (1997). Requirement for engrailed and invected genes reveals novel
regulatory interactions between engrailed/invected, patched, gooseberry and wingless during
Drosophila neurogenesis. Development 124: 1675-1688. Medline abstract: 97307774
Brody, T. and Odenwald, W. F. (2000). Programmed transformations in neuroblast gene expression during Drosophila CNS lineage development. Dev. Biol. 226(1):34-44. Medline abstract: 10993672
Campos-Ortega, J.A. (1995). Genetic mechanisms of early neurogenesis in Drosophila
melanogaster. Mol. Neurobiol. 10: 75-89. Medline abstract: 96063027
Chu-LaGraff, Q. and C.Q. Doe. (1993). Neuroblast specification and formation regulated by
wingless in the Drosophila CNS. Science 261: 1594-1597. Medline abstract: 93383272
Cui X. and Doe C. Q. (1995) The role of the cell cycle and cytokinesis in regulating neuroblast sublineage gene expression in the Drosophila CNS. Development, 121: 3233-3243. Medline abstract: 7588058
D'Alessio, M. and M. Frasch. (1996). msh may play a conserved role in dorsoventral patterning
of the neuroectoderm and mesoderm. Mech. Dev. 58: 217-231. Medline abstract: 97042058
Harris, W. A. (1997). Cellular diversification in the vertebrate retina. Curr. Opin. Genet. Dev. 7: 651-658. Medline abstract: 9388782
Harris, W. A. (2001). Temporal coordinates: the genes that fix cell fate with birth order. Developmental Cell 1: 313-314
Isshiki, T., M. Takeichi, and A. Nose. (1997). The role of the msh homeobox gene during
Drosophila neurogenesis: Implication for the dorsoventral specification of the neuroectoderm.
Development 124: 3099-3109. Medline abstract: 97417578
Jimenez, F., L.E. Martin-Morris, L. Velasco, H. Chu, J. Sierra, D.R. Rosen, and K. White.
(1995). vnd, a gene required for early neurogenesis of Drosophila, encodes a homeodomain
protein. EMBO J. 14: 3487-3495. Medline abstract: 95354667
Kambadur, R., et al., (1998). Regulation of POU genes by castor and hunchback establishes layered compartments in the Drosophila CNS. Genes Dev. 12(2): 246-260. Medline abstract: 98099749
Landgraf, M., et al. (2003). Charting the Drosophila neuropile: a strategy for the standardised characterisation of genetically amenable neurites. Dev. Bio. 260: 207-225. Medline abstract: 12885565
Matsuzaki, M. and K. Saigo. (1996). hedgehog signaling independent of engrailed and wingless required for post-S1 neuroblast formation in Drosophila CNS. Development 122: 3567-3575. Medline abstract: 97108751
McDonald, J.A. and C.Q. Doe. (1997). Establishing neuroblast-specific gene expression in the Drosophila CNS: Huckebein is activated by Wingless and Hedgehog and repressed by Engrailed and Gooseberry. Development 124: 1079-1087. Medline abstract: 97209482
McDonald, J.A., S. Holbrook, T. Isshiki, J. Weiss, C.Q. Doe, and D.M. Mellerick. (1998). Dorsoventral patterning in the Drosophila CNS: The vnd homeobox gene specifies
ventral column identity. Genes Dev. 12: 3603-12. Medline abstract: 99051322
McConnell, S. K. (1995). Constructing the cerebral cortex: neurogenesis and fate determination. Neuron 15: 761-768. Medline abstract: 7576626
Mellerick, D.M. and M. Nirenberg. (1995). Dorsal-ventral patterning genes restrict NK-2 homeobox gene expression to the ventral half of the central nervous system of Drosophila embryos. Dev. Biol. 171: 306-316. Medline abstract: 96004999
Raz, E. and B.Z. Shilo. (1993). Establishment of ventral cell fates in the Drosophila embryonic ectoderm requires DER, the EGF receptor homolog. Genes Dev. 7: 1937-1948. Medline abstract: 94010291
Rutledge, B.J., K. Zhang, E. Bier, Y.N. Jan, and N. Perrimon. (1992). The Drosophila spitz
gene encodes a putative EGF-like growth factor involved in dorsal-ventral axis formation and
neurogenesis. Genes Dev. 6: 1503-1517. Medline abstract: 92354912
Schweitzer, R., M. Shaharabany, R. Seger, and B.Z. Shilo. (1995). Secreted Spitz triggers the
DER signaling pathway and is a limiting component in embryonic ventral ectoderm
determination. Genes Dev. 9: 1518-1529. Medline abstract: 95324812
Skeath, J.B., G.F. Panganiban, and S.B. Carroll. (1994). The ventral nervous system defective
gene controls proneural gene expression at two distinct steps during neuroblast formation in
Drosophila. Development 120: 1517-1524. Medline abstract: 94326649
Skeath, J.B., Y. Zhang, R. Holmgren, S.B. Carroll, and C.Q. Doe. (1995). Specification of
neuroblast identity in the Drosophila embryonic central nervous system by gooseberry-distal.
Nature 376: 427-430. Medline abstract: 95356833
Skeath, J. B. (1998). The Drosophila EGF receptor controls the formation and specification of
neuroblasts along the dorsal-ventral axis of the Drosophila embryo. Development 125(17):
3301-3312. Medline abstract: 98359942
Truman, J. W. and Ball, E. E. (1998). Patterns of embryonic neurogenesis in a primitive wingless insect,
the silverfish, Ctenolepisma longicaudata: comparison with those
seen in flying insects. Dev. Genes Evol. 208(7): 357-68. Medline abstract: 98404279
Udolph G., et al. (1998). Differential effects of EGF receptor signalling on neuroblast lineages along the dorsoventral axis of the Drosophila CNS. Development 125(17): 3291-3299. Medline abstract: 98359941
Weigmann K. and Lehner C. F. (1995) Cell fate specification by even-skipped expression in the Drosophila nervous system is coupled to cell cycle progression. Development. 121: 3713-3721. Medline abstract: 8582283
Weiss, J. B. Ohlen, T. V. Mellerick, D. M., Dressler, G. Doe, C. Q. and Scott, M. P. (1998). Dorsoventral patterning in the Drosophila central nervous system:
the intermediate neuroblasts defective homeobox gene specifies
intermediate column identity. Genes Dev 12: 3591-3602. Medline abstract: 99051321
White, K., N.L. DeCelles, and T.C. Enlow. (1983). Genetic and developmental analysis of the
locus vnd in Drosophila melanogaster. Genetics 104: 433-448. Medline abstract: 83288499
Yagi, Y., Suzuki, T. and Hayashi, S. (1998). Interaction between Drosophila EGF receptor and
vnd determines three dorsoventral domains of the neuroectoderm. Development 125(18):
3625-3633. Medline abstract: 98384319
Zhang, Y., A. Ungar, C. Fresquez, and R. Holmgren. (1994). Ectopic expression of either the
Drosophila gooseberry-distal or proximal gene causes alterations of cell fate in the epidermis
and central nervous system. Development 120: 1151-1161. Medline abstract: 94298517
Genes involved in organ development
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
Please e-mail comments/corrections to brodyt@codon.nih.gov