BIOLOGICAL OVERVIEW

The Drosophila bHLH transcription factor target of poxn (tap; also termed biparous) is expressed in a small subset of neurons when they undergo differentiation. Although the sequence of tap resembles that of proneural genes, in fact it is a neural differentiation gene, acting later in the hierarchy of gene activation than the proneural genes (Bush, 1996 and Gautier, 1997). In the CNS tap is expressed in late neural precursors prior to their exit from the the mitotic cycle, but tap is not expressed in neurons (Bush, 1996). In the peripheral nervous system, tap is expressed exclusively in one of the neurons that innervate each larval chemosensory organ, possibly controlling the specific properties of that neuron. Sequence comparisons suggest that tap is most closely related to two bHLH genes, neurogenin and neuroD, identified in several vertebrate species; these genes are involved respectively in neural determination and in neuronal differentiation (Gautier, 1997). tap is also expressed at a late stage in the development of the gustatory (chemosensory) bristles of the leg, wing and proboscis. tap is expressed very early in the development of a second type of chemosensory receptors, the olfactory organs of the antenna (Ledent, 1998).

Numerous vertebrate homologs of Drosophila proneural genes of the achaete-scute complex (AS-C) and of atonal have been cloned, supporting the argument that many aspects of neurogenesis in Drosophila and vertebrates are homologous. Indeed, characteristics consistent with a role in neural precursor determination have been demonstrated for some of these genes, including Xash3 (a Xenopus AS-C homolog); Math1 and Math5 (ato homologs), and neurogenin1 (a tap homolog). Surprising was the fact that expression of many of these homologs are activated later in neural development, after the stage of neural precursor determination (Golding, 2000).

Examples of such genes that play late roles in neurogenesis included asense (ase) an aberrant member of the AS-C, and cato, a gene that bears close relationship to proneural gene atonal. Mutation of ase has no apparent effect in most of the cells in which it is expressed, which may be because the phenotype is too subtle or there is redundancy with other factors. cato is expressed widely in the developing PNS after neural precursor selection but before terminal differentiation. Consistent with this pattern, cato appears to be required for proper sensory neuron morphology. It is clear that cato is not a proneural gene, despite its close sequence relationship with ato. The available evidence suggests that instead, it is associated later in neurogenesis with correct neuronal morphology. cato therefore potentially represents a gene that functions similar to Drosophila tap and vertebrate neuronal differentiation regulators, such as neuroD (Golding, 2000).

From these observations, it has been proposed that cascades or networks of bHLH factors may function during all stages of neural development from commitment of precursors, through to proliferation and migration, and finally to postmitotic terminal differentiation. The implication of this hypothesis is that bHLH proteins control appropriate target genes for each of these stages, including the activation of the next bHLH protein in the cascade (Golding, 2000b and references therein).

tap is a potential target of Paired box neuro (Pox-n). pox-n is expressed in two clusters of cells in each segment, one dorsal and one ventral. The dorsal-most cluster is displaced laterally in the second and third thoracic segments, a pattern typical of the chemosensory organs. The expression of tap follows the same pattern, with two important differences. (1) While pox-n is expressed in the sensory mother cell (SMC) and throughout the lineage until shortly before the progeny undergo differentiation, tap is expressed only at or near the onset of differentiation. Thus tap expression is very transient, lasting probably for less than an hour. (2) While pox-n is expressed in most or all of the progeny of the SMC, tap is expressed in only one cell of each organ (Gautier, 1997).

It has been confirmed that tap depends (directly or indirectly) on pox-n by inducing the ectopic expression of pox-n early during embryogenesis. the overexpression of pox-n has been shown to result in the development of ectopic chemosensory organs, both in the larva and in the adult. Additional cells expressing tap were observed embryos were the ectopic expression of pox-n was induced at 4-6 h after egg laying. Conversely, in embryos homozygous for a deficiency removing pox-n, the expression of tap is completely abolished and largely but not completely so in the CNS. Pox-n binds to polytene chromosomes at 74B, the same location that codes for tap (Gautier, 1997).

tap is a perfect example of a neural precursor gene, silent during the early stages of neurogenesis but activated during the generation of neural sublineages. The transient expression of tap in GMCs also demonstrates that these immediate progeny of the neuroblast have their own program of gene expression, thus pointing to the necessity for more intensive investigations of GMC gene expression. Regulated gene expression in the GMC represents an additional level of complexity in the process of neurogenesis, beyond the dynamic expression of genes in the neuroblast.

To determine the expression pattern of TAP mRNA during embryogenesis, in situ hybridization experiments were performed. TAP mRNA is expressed in a dynamic and complex pattern, starting at stage 10 and lasting beyond stage 15. An important feature of Tap expression at all stages of embryogenesis is its brief duration in select groups of cells. The earliest expression is seen at stage 10 in 5 cells per segment: an unpaired cell on the midline and two pairs lateral to the midline. Based on their large size and position in the neuroblast layer, the midline cell and the pair of cells just lateral to the midline are neuroblasts. Double-labeling of embryos for the expression of the Engrailed protein and TAP mRNA reveals that the position of the cell on the midline coincides with an En stripe, suggesting that this cell is the median neuroblast. Double-labeling of embryos for the expression of the wingless-lacZ gene and TAP mRNA reveals that the more lateral of the paired cells is a row 4 neuroblast, residing one cell anterior to the row 5 neuroblasts which express wingless. Its position and lateral shape suggest that it is neuroblast 4-1. During embryogenesis, both the median neuroblast and neuroblast 4-1 generate solely neuronal progeny. During germ-band extension, there is a large burst of biparous expression in clusters of cells in the ganglion mother cell layer of the developing ventral nerve cord. The cells expressing TAP at this time are distinguishable from neuroblasts based on their location and their smaller size. Although precise quantitation is difficult, it is estimated that 12-18 cells per hemisegment in the ganglion layer express Tap between 6 and 7 hr of development. After this large burst of expression, TAP mRNA is restricted to progressively fewer cells. At embryonic stage 14, TAP mRNA is found in cells peripheral to the longitudinal axon tracts, at the lateral edge of the CNS. By stage 15, 2 large cells per hemisegment maintain tap expression (Bush, 1996).

To determine whether TAP-positive cells are neurons, embryos were double-labeled for the ELAV protein and the TAP transcript. ELAV is an RNA binding protein expressed in all postmitotic neurons. These experiments reveal that TAP mRNA is largely excluded from ELAV-positive cells. This is true even at stage 15 when the vast majority of cells in the CNS have reached the postmitotic state. Further confirmation of the absence of TAP staining from neurons was obtained using anti-HRP antiserum, which recognizes all neurons. To determine whether TAP-expressing cells express the glial-specific gene repo, embryos were double-labeled for the TAP transcript and the REPO protein. Cells express repo after they have committed to the glial cell fate. The large burst of tap expression in the ganglion mother cell layer precedes most REPO expression (stage 11 vs stage 12). Later in neurogenesis when the two genes do overlap, most TAP-positive cells are REPO negative. However, close inspection of these embryos shows that a TAP-positive cell is sometimes directly apposed to a REPO-positive cell. Therefore, the possibility that in some cases expression of the two genes might overlap for a very short time in the same cell cannot be excluded. Since biparous-positive cells express neither elav nor repo, it is possible that they are not yet postmitotic. To determine whether TAP-positive cells are synthesizing DNA, embryos were labeled with bromodeoxyuridine (BrdU). Some TAP-positive cells also incorporate BrdU. This demonstrates that TAP-positive cells are synthesizing DNA, suggesting that they are progenitor cells that give rise to neurons and/or glia (Bush, 1996).

To follow the fate of the TAP-expressing cells, a fusion gene was constructed in which 5.5 kb of the tap promoter drives expression of lacZ. It was reasoned that perdurance of beta-galactosidase would permit observation of cells that had turned on biparous even after disappearance of the transcript. The promoter fragment used in this experiment included 166 N-terminal amino acids of Tap. Staining of these embryos for beta-galactosidase revealed a recapitulation of part of the late expression pattern of endogenous tap. Specifically, beta-galactosidase immunoreactivity was observed in the position of TAP-positive cells at stages 12 through 15, with other phases of biparous expression absent. Similar to endogenous tap, the expression of this transgene is transient. This suggests that the tap promoter/lacZ transgene possesses sufficient cis-regulatory information to direct beta-galactosidase expression to TAP-positive cells at stage 12 through 15 (Bush, 1996).

To confirm that the beta-galactosidase-positive cells recapitulate the expression of tap, embryos were double-labeled for beta-galactosidase and the TAP transcript. If the fusion gene reflects the endogenous TAP pattern, one would predict that some cells would express both TAP transcript and beta-galactosidase protein as was observed. In addition, beta-galactosidase appears to persist beyond the time when the TAP transcript is no longer detectable by in situ hybridization. The beta-galactosidase signal is observed in cells that have migrated both medially and laterally (and this is probably a result of the perdurance of the beta-galactosidase protein. Alternatively, the lacZ fusion gene could be missing cis-elements necessary for turning off tap expression and therefore lacZ is still transcribed after the endogenous gene has been shut off. This possibility is thought to be unlikely since beta-galactosidase expression is only slightly more prolonged than TAP mRNA. Finally, this expression pattern has been observed with at least five independent insertions, demonstrating that the regulatory information contained within the promoter is independent of chromosomal location (Bush, 1996).

It was also observed that the anti-beta-galactosidase staining is nuclear. Since the lacZ gene used in the biparous promoter construct encodes a cytoplasmic protein, this result indicates that the N-terminal portion of Biparous is able to direct nuclear localization. Ten residues N-terminal to the fusion point is a sequence, KRFRR, which conforms well to the canonical nuclear localization sequence, KKRK, suggesting that this sequence may be responsible for nuclear localization. Since the beta-galactosidase signal persisted longer than the TAP transcript, an examination was made to see whether the beta-galactosidase-positive cells express either neuronal or glial markers. When stage 12 embryos were costained with antisera to REPO and beta-galactosidase, some cells were clearly double-labeled. In this region, repo is expressed in the segmental nerve glia, the intersegmental nerve glia, and the exit glia. These three sets of glia are all lateral to the longitudinal connectives, with the segmental and intersegmental glia associated with their respective nerve tracts and the exit glia located at the boundary between the CNS and PNS. Since the double-labeled cells may have not reached their final destination, it is not evident which of these three groups of glia are double-labeled. Nevertheless, since repo is a glial-specific gene, the colocalization provides evidence that the tap gene is transcribed in precursor cells that generate glial progeny. To determine whether beta-galactosidase protein is also expressed in neurons, embryos were double labelled for ELAV and beta-galactosidase. Some cells simultaneously express both proteins. This shows that the tap promoter is also activated in neuronal precursors (Bush, 1996).

Since tap is expressed in glial precursors, it was of interest to determine its relationship to gcm, a gene required for normal glial development. Both biparous and gcm are expressed relatively early during CNS development, with both transcription factors, turned on before many neuronal and glial progenitors have completed their final mitosis. Nevertheless, it is important to note that the large burst of TAP mRNA expression in the ganglion mother cell layer precedes most gcm expression (stage 11 vs stage 12). This sequence of expression is consistent with tap being upstream of gcm in gliogenesis. In support of this hypothesis, the pattern of biparous expression is unchanged in gcm mutant embryos. Since TAP mRNA is expressed relatively early in neurogenesis, it is also possible that it is a downstream target of the proneural genes of the achaete-scute complex. To determine the effect of these genes on biparous expression, in situ hybridization experiments were performed with achaete-scute mutant backgrounds. It was found that removing asense or achaete and scute has no effect on the pattern of TAP mRNA (Bush, 1996).

One theory suggests that bHLH genes determine primordial cell types early in development, and this is followed by the action of other genes that influence the appearance of more specialized cell types This holds true for Drosophila where the bHLH genes of the achaete-scute complex confer on neuroectodermal precursor cells the ability to become neuroblasts. Although tap is activated in neuroblasts, its expression persists until the final stages of embryogenesis, considerably after the disappearance of achaete, scute, and lethal of scute. Thus, tap may present a new function for neuronal bHLH genes, since it is expressed at a different time and in fewer cells than the achaete-scute complex genes. Given its late and more restricted expression pattern, one potential role for tap would be to specify unique characteristics of individual lineages of neurons and glia (Bush, 1996).

GENE STRUCTURE

cDNA clone length - 1.5 kb

PROTEIN STRUCTURE

Amino Acids - 398

Structural Domains

The bHLH motif of Tap is closest to those coded for by mouse neurogenin and Xenopus neurogenin-related 1A (75% identity over the entire bHLH motif), and slightly more distant to the Xenopus and mouse neuroD genes (62% identity). This level of identity between Tap and Neurogenin is higher than that observed between the products of the achaete-scute complex and their MASH relatives, which ranges around 65%. The identity of the bHLH domain of Tap with its closest relative among the fly proneural proteins, Atonal, is only 50%. Furthermore, a detailed analysis reveals significant residue identities in the basic regions of Tap and its Neurogenin/NeuroD relatives. Thus tap is clearly more similar to the neurogenin/neuroD subfamily, than to the other fly (or vertebrate) neural bHLH genes (Gautier, 1997).

Three mRNA instability motifs are found in the 3' untranslated region of TAP mRNA, suggesting that the TAP mRNA might be short-lived, an hypothesis supported by expression studies. The predicted Biparous protein has 407 amino acids, with the bHLH domain in the center of the protein. Thirty residues C-terminal to the bHLH domain, amino acids 244-333, is a region in which 38% of the residues are either proline or glutamine. Proline/glutamine-rich regions define a major class of transcription activation domains. In addition, C-terminal to the potential activation domain, starting at amino acid 349, is a region highly homologous to the NH (Numb homology) motif of the homeobox transcription factor Prospero. Within this motif Prospero is more similar to Biparous than to Numb. Further experiments will be necessary to determine if Biparous is asymmetrically segregated (Bush, 1996).

date revised: 3 December 2000

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.