BIOLOGICAL OVERVIEW

Steroid hormones coordinate multiple cellular changes, yet the mechanisms by which these systemic signals are refined into stage- and tissue-specific responses remain poorly understood. The Drosophila gene Eip93F, more familiarly termed E93 determines the nature of a steroid-induced biological response. E93 mutants possess larval salivary glands that fail to undergo steroid-triggered programmed cell death, and E93 is expressed in cells immediately before the onset of death. E93 protein is bound to the sites of steroid-regulated and cell death genes on polytene chromosomes, and the expression of these genes is defective in E93 mutants. Furthermore, expression of E93 is sufficient to induce programmed cell death. It is proposed that the steroid induction of E93 determines a programmed cell death response during development (Lee, 2000).

The mechanisms of steroid signaling have been extensively studied in Drosophila larval salivary glands by virtue of the giant polytene chromosomes that form ecdysone-induced puffs, reflecting a transcriptional regulatory hierarchy. The ecdysone receptor complex, a heterodimer of the EcR and Usp nuclear receptors, activates transcription of a small set of early regulatory genes. These genes encode transcription factors that, in turn, activate a larger set of late genes, which are thought to play a more direct role in controlling the appropriate biological responses to the hormone. Previous studies have implicated the EcR, usp, βFTZ-F1, BR-C, and E74A genes in steroid-activated larval cell death. The role of βFTZ-F1 appears to be indirect, functioning as general competence factor for prepupal responses to ecdysone. In contrast, EcR, Usp, BR-C, and E74A play a more direct role in triggering salivary gland cell death through the coordinate induction of rpr and hid transcription. These factors, however, are not sufficient for the death response because they do not direct this pathway in response to the earlier pulse of ecdysone at puparium formation. Rather, one or more stage-specific regulators must be induced by ecdysone at the end of prepupal development that determine the stage specificity of salivary gland cell death. The E93 early gene is an ideal candidate for fulfilling this function. E93 is induced as a primary reponse to ecdysone in a stage- and tissue-specific manner (Baehrecke, 1995). E93 transcription increases immediately prior to larval midgut and salivary gland cell death and is coordinately induced with rpr and hid (Baehrecke, 1995; Jiang, 1997). This correlation suggests that E93 may contribute to the stage specificity of larval tissue cell death (Lee, 2000).

The expression of E93 protein in dying cells, combined with the defects in E93 mutant salivary gland cell death and transcription of apoptosis genes, indicates that E93 is a key determinant of steroid-induced programmed cell death. Therefore, tests were performed to see if expression of E93 is sufficient to kill wing imaginal disc cells that have a well-defined response to ecdysone during metamorphosis. This experiment was carried out by crossing UAS-E93 transformant flies with Drosophila strains that express GAL4 in wing imaginal discs (vg-GAL4). All progeny that possess both UAS-E93 and vg-GAL4 die at the start of pupal development. This lethal phase, combined with the wealth of information about ecdysone-triggered wing development led to a detailed characterization of E93-induced cell death in vg-GAL4 UAS-E93 animals. Control (UAS-E93 alone or vg-GAL4 alone) wing-thoracic imaginal discs dissected 2 hr following puparium formation exhibit little cell death. By contrast, E93-expressing wing-thoracic imaginal discs, exhibit extensive cell death in the wing blade and hinge regions at this developmental stage. This pattern of cell death mimics the pattern of GAL4 expression in this vg-GAL4 strain of Drosophila, as determined by crossing vg-GAL4 flies with a UAS-lacZ reporter and detecting β-galactosidase activity. Animals expressing E93 under the control of vg-GAL4 die soon after head eversion during metamorphosis and exhibit defects in the presumptive adult notum and wing. These animals were examined further by characterizing the morphology of wing imaginal discs prior to the lethal phase but after the induction of ectopic cell death. While control wing-thoracic imaginal discs dissected from animals 4 hr following puparium formation have clearly progressed in elongation of the wing, E93-expressing animals of the same age possess defective wing-thoracic imaginal discs that do not properly elongate. These data demonstrate that expression of E93 is sufficient to induce programmed cell death (Lee, 2000).

It is concluded that the precise temporal and spatial patterns of E93 induction by the steroid hormone ecdysone determines the biological fate of those target tissues, directing the massive programmed cell death of the larval salivary glands during metamorphosis. E93 also acts both directly and indirectly to regulate the transcription of key effector genes that drive the cell death response. Initial studies of the ecdysone-triggered gene cascades speculated that early ecdysone-induced regulatory genes might be expressed in a tissue-specific manner, directing the different fates of larval and adult cells during metamorphosis. In contrast, molecular characterization of the BR-C, E74, and E75 early genes demonstrates that these genes are widely expressed throughout the animal. Localization of EcR and BR-C protein isoforms reveals that they are expressed in subsets of ecdysone target tissues; however, these expression patterns do not correlate with sets of tissues that undergo one fate in response to ecdysone. Similarly, studies of EcR, usp, BR-C, and E74 mutants have revealed multiple functions for these genes, affecting the development of both larval and adult cells during metamorphosis. These observations led to the 'tissue coordination model', which proposes that overlapping combinations of early ecdysone-induced transcription factors dictate the proper tissue-specific responses to ecdysone pulses during development (Lee, 2000).

E93 stands in sharp contrast to these widely expressed early genes. E93 expression is restricted to metamorphosis, with induction in the midguts of newly formed prepupae preceding induction in the salivary glands of late prepupae. The temporal correlation of this expression pattern with the onset of midgut and salivary gland cell death raised the possibility that E93 might play a role in regulating the death response (Baehrecke, 1995). Strong evidence in support of this model is proved. Antibody stains show that E93 is induced in a cell type-specific pattern in the larval midguts, restricted to the polytene larval cells that are fated to die, and excluded from the diploid imaginal cells that form the adult gut. Similarly, E93 protein expression in the salivary gland parallels that of its mRNA, immediately preceding cell death. E93 mutants die as pupae with persistent salivary glands, and this salivary gland cell death defect can be rescued by E93 expression from a transgene. Moreover, ectopic E93 expression is sufficient to direct a death response. Thus, the ecdysone induction of E93 defines the fate of that tissue, directing its immediate and massive destruction through programmed cell death. E93 regulation therefore provides a molecular mechanism for refining the systemic ecdysone signal into a specific biological response during development (Lee, 2000).

Although E93 encodes a novel protein with little similarity to other proteins in the sequence databases, it shares several characteristics of Drosophila transcription factors. These include homopolymeric tracts of acidic amino acids that can serve as transcriptional activation domains, a potential nuclear localization signal, and two basic domains that could serve as DNA binding motifs. E93 is localized to nuclei and binds to specific sites on polytene chromosomes, further suggesting that E93 regulates gene activity. Significantly E93 mutants impact the transcription of genes from chromosome loci that are bound by E93 protein. While these data do not provide conclusive evidence that distinguish the biochemical characteristics of DNA and chromatin binding, these results are consistent with the hypothesis that E93 encodes a novel transcription regulator (Lee, 2000).

E93 appears to exert its effects by both directly and indirectly regulating genes required for programmed cell death. Prior to the prepupal stage, ecdysone triggers regulatory hierarchies that do not result in salivary gland cell death. In late third instar larvae, for example, the ecdysone receptor complex activates the primary response genes BR-C, E74, and E75. These early genes, in turn, direct a switch in salivary gland late gene expression, repressing the glue genes and inducing more than 100 late genes including the L71 genes. The following pulse of ecdysone, at the end of prepupal development, triggers the BR-C, E74, E75, and E93 early genes. This response is dependent on the prior expression of the βFTZ-F1 competence factor, which is necessary and sufficient for early gene induction in late prepupae. As expected βFTZ-F1 is expressed at a normal level in E93 mutants with a delay due to genetic background. In contrast E93 is required for ecdysone induction of the BR-C, E74A, and E75A genes in prepupal salivary glands. E93 mutants do not impact EcR and E74B transcription and head eversion, indicating that the prepupal pulse of ecdysone is normal in these animals. Thus, the effect of E93 mutants on transcription of early genes is not caused by the absence of ecdysone. E93 protein binds to the 74EF and 75B puffs that contain the E74 and E75 genes, suggesting that these are direct regulatory targets. E93 does not bind to the 2B5 puff containing the BR-C gene, suggesting that this regulation is indirect (Lee, 2000).

Several cell death genes are transcribed immediately prior to larval salivary gland programmed cell death. Components of the core apoptosis machinery, including Ark and the caspase dronc, as well as the death genes rpr, hid, and crq, increase in transcription in late prepupal salivary glands. The synchronous induction of these cell death genes indicates that salivary glands die by a mechanism that is similar to that utilized in apoptosis during Drosophila embryogenesis, where rpr and hid are involved in caspase activation. The increase in crq transcription in dying salivary glands suggests that these cells are unique, however, since Crq is expressed in phagocytes and functions in removal of dying cells during embryogenesis. This increase in crq transcription is not due to the adhesion of phagocytes to the dying salivary gland, since Crq protein is expressed at a high level in the dying cells. While salivary gland cell destruction involves genes that function in apoptosis, these cells also have characteristics of autophagy, and this form of cell death may utilize crq in the terminal stages of cell removal (Lee, 2000).

E93 mutants impact transcription of cell death genes, consistent with the model that E93 serves as a regulator that specifies the cell death response to ecdysone. E93 mutants exhibit defects in transcription of rpr, hid, crq, ark, and dronc. The 75C locus that contains rpr and hid and the 53F locus that contains ark are not bound by E93 protein, suggesting that the regulation of these genes is indirect. Recent studies of rpr and hid regulation support this conclusion, since mutations in BR-C and E74A alter rpr and hid RNA levels, and E93 is required for BR-C and E74A expression. The 21C locus, which contains crq, is bound by E93 protein, suggesting that E93 may directly regulate crq transcription. Thus, E93 plays an essential role in regulating cell death genes, thereby directing steroid-triggered programmed cell death (Lee, 2000).

This study demonstrates that components of the central cell death pathway, including Ark and dronc, exhibit dynamic changes in RNA transcription that immediately precede salivary gland cell death. It is important to consider that many of these factors may also be regulated at the posttranscriptional level. For example, rpr and hid direct programmed cell death during Drosophila embryogenesis by repressing the inhibitory activity of DIAP1 on caspase activation. Thus, while rpr, hid, and dronc are regulated by the ecdysone-induced primary response genes at the transcriptional level, dronc may also be regulated by a secondary mechanism. Future studies of the genetic pathways that mediate steroid-regulated destruction of larval salivary glands will provide further insights into the conserved molecular mechanisms that underlie cell death during development (Lee, 2000).

GENE STRUCTURE

cDNA clone length - 9567

Bases in 5' UTR - 393

Exons - 5

Bases in 3' UTR - 5508

PROTEIN STRUCTURE

Amino Acids - 1221

Structural Domains

E93 spans at least 55 kb of genomic DNA and encodes a 146-kDa protein that has no matches in the sequence databases, but displays several characteristics of known Drosophila transcription factors. It is proposed that E93 acts in a stage-specific regulatory hierarchy in the salivary gland to direct its histolysis in response to the prepupal ecdysteroid pulse (Baehrencke, 1995).

Many prokaryotic and eukaryotic DNA-binding proteins use a helix-turn-helix (HTH) structure for DNA recognition. A new family of eukaryotic HTH proteins, the Pipsqueak (Psq) family, is described that includes proteins from fungi, sea urchins, nematodes, insects, and vertebrates. Three subgroups of the Psq family can be distinguished. Like the HTH proteins of the prokaryotic resolvase family, members of the CENP-B/transposase subgroup catalyze site-specific recombination reactions. This functional conservation, together with a primary sequence similarity between the resolvase and Psq DNA-binding domains, suggests that the resolvase and Psq families are evolutionarily linked. More than half of the newly identified Drosophila Psq proteins contain a BTB protein-protein interaction domain. All proteins of this BTB subgroup belong to the conserved Tramtrack group of BTB-domain proteins. About half of the members of the Tramtrack group contain a Psq domain, while the other half is made up of proteins that contain a zinc finger domain. Thus, nearly all members of this group appear to be DNA-binding proteins. Among other developmental regulators, the Drosophila cell death protein E93 was found to contain a Psq motif and to define a third subgroup of Psq domain proteins. The high sequence conservation of the E93 Psq motif allowed the identification of E93 orthologs in humans and lower metazoans (Siegmund, 2002).

EVOLUTIONARY HOMOLOGS

Mushroom bodies (MBs) are considered to be involved in higher-order sensory processing in the insect brain. To identify the genes involved in the intrinsic function of the honeybee MBs, genes preferentially expressed therein were sought using the differential display method. A novel gene encoding a putative transcription factor (Mblk-1) is expressed preferentially in one of two types of intrinsic MB neurons -- the large-type Kenyon cells. Mblk-1 is thus a candidate gene involved in the advanced behaviours of honeybees. A putative DNA binding motif of Mblk-1 has significant sequence homology with those encoded by genes from various animal species, suggesting that the functions of these proteins in neural cells are conserved among the animal kingdom (Takeuchi, 2001).

The Mblk-1 gene in the honeybee brain encodes a transcription factor containing two DNA binding motifs, termed RHF1 and 2. Two mouse Mblk1 homologs, Mlr1 and Mlr2, have been identified. Both encode proteins containing a single DNA-binding motif highly conserved with RHF2, and both activate transcription mediated by a DNA element recognized by honeybee Mblk-1. Mlr1 is expressed predominantly in the spermatocytes of the testis, while Mlr2 is expressed in various tissues other than testis. Mlr1 transcripts are lost in the testis of W/W(v) mutant mice, suggesting a role in spermatogenesis (Kunieda, 2003).

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.