Gene name - Ecdysone-induced protein 63E
Synonyms - L63
Cytological map position - 63E1-4
Function - signal transduction
Keywords - molting
Symbol - Eip63E
FlyBase ID: FBgn0005640
Genetic map position - 3-
Classification - cyclin dependent kinase
Cellular location - cytoplasm
The puffing response to the late-larval ecdysone pulse consists in the rapid induction of six early puffs in a primary response to the hormone, followed by a secondary response consisting in the regression of these early puffs and the concomitant induction of over 100 late puffs. It is thought that the early puffs represent the expression of genes involved in the regulation of the molting heirarchy, and that the late puffs represent the expression of effector genes. The L63 gene is complex: it codes for Ecdysone-induced protein 63E (Eip63E), which is responsible for the late larval 63E puff. It spans 85 kb of genomic DNA that includes three overlapping transcription units that generate nine different mRNAs encoding five different L63 isoforms. In these regards, L63 more closely resembles the BR-C, E74, and E75 early genes than it does the simple Sgs and L71 secondary-response genes. L63 also differs from these genes in its temporal and spatial expression patterns, again more closely resembling the early genes. Whereas Sgs and L71 gene expression is limited to the terminal stages of the larval salivary gland's existence, one or more of the L63 transcription units are expressed throughout Drosophila development and in many different tissues. These characteristics provide for a redundancy of expression in a gene that mutational analysis shows is required even before metamorphosis occurs. The L63 isoforms contain a common 294-amino-acid region near the carboxy-terminus that shows strong homology to the cyclin-dependent kinase (CDK) protein family. This region provides a CDK function critical to development (Stowers, 2000).
It has been argued that the L63 proteins function as cyclin-dependent kinases, given the strong sequence similarity of L63 to established CDKs, the observation that a mutant which removes a key catalytic site is a likely null, and the results of the rescue experiments with altered L63 proteins generated of in vitro mutagenesis. However, there is no evidence that L63 is involved in the control of cell division as is the case for many CDKs. In fact, several observations suggest otherwise. (1) Qualitative examination of the brains of wild-type and deficiency larvae at pupariation revealed no significant difference in the fraction of cells undergoing mitosis. (2) No differences in the deficiency of polytenization in the salivary gland polytene chromosomes are seen in wandering third-instar mutant larvae, when compared to wild-type animals. (4) Eyes composed entirely of L63 clones are indistinguishable from control mitotic recombinant eyes, indicating that L63 is not required for cell division in the eye. These results suggest that the CDK functions of the L63 proteins are required for functions other than the regulation of cell division. A similar proposal has been made for the closest mammalian relatives of L63 -- namely, the PFTAIRE and PCTAIRE proteins, which may play a role in cell differentiation rather than in cell division (Stowers, 2000).
Given the above observations, the wide spatial and temporal distributions of the L63 proteins, the functional redundancies indicated by the L63 mutants, and the observation that L63 overexpression appears to have no phenotypic consequences, it is proposed that the functional specificity of L63 proteins derives in large measure from associated cyclins and/or other proteins that regulate L63 activity. For instance, an interaction between L63 and one cyclin in an imaginal tissue, like the leg disc (whose metamorphic fate is to form an adult structure) could result in L63 controlling epithelial morphogenesis, while an interaction between L63 and a different cyclin in a strictly larval tissue (such as the salivary gland whose fate during metamorphosis is histolysis) could result in L63 playing a role in programmed cell death. This scenario is consistent with the functions observed for other CDKs. Such a functional specificity has been observed for the mammalian CDK5, a key regulator of neuronal differentiation (see Drosophila CDK5). CDK5 is widely expressed in human and mouse tissues, but its kinase activity is restricted to the brain where its specific activator p35/p25 is expressed. Another example of specific control of CDK activity by different cyclins is that of the S. cerevisiae PHO85 CDK protein. It is essential for the proper regulation of phosphate and glycogen metabolism as well as for cytokinesis. To accomplish these roles it appears that PHO85 may well interact with as many as 10 different cyclins. Mutation of some of these cyclins has shown that they can produce mutant phenotypes that are a subset of those of the PHO85 mutant. Apparently these cyclins exhibit overlapping but unique functions. Using these results as a model, it can be imagined that L63 proteins interact with quite different cyclins or other proteins to provide different developmental functions (Stowers, 2000 and references therein).
Evidence that L63 is required for embryogenesis consists in the observation that L63 mutant embryos that are maternally and zygotically deficient for L63 activity do not complete embryogenesis. Interestingly, recent observations make the possibility plausible that L63 maternal expression may be ecdysone regulated during oogenesis because the same set of early genes induced by ecdysone at the onset of metamorphosis is coordinately regulated by ecdysone during oogenesis. Maternal L63 expression may therefore be part of an ecdysone-regulated hierarchy of gene expression in oogenesis that is used repeatedly during development (Stowers, 2000).
The observation that the majority of zygotically deficient L63 mutants die as larvae indicates that L63 is also required during larval development. From the small pupal phenotype of the few L63 mutants that survive to pupariation it is inferred that L63 may participate in the regulation of organism size. There are numerous ways in which the loss of L63 activity could indirectly reduce organism size; these include, to name a few, a role for L63 proteins in digestion, muscle contraction, or the nervous system. An interesting possibility is that L63 proteins act as direct determinants of organism size. One possible scenario is that these proteins are part of a system that regulates progression from one developmental stage to the next, so that when L63 is absent, precocious developmental transitions occur (Stowers, 2000).
The bent-leg phenotype of rare escaper adult L63 mutants indicates a third function of L63 in development: an involvement in epithelial morphogenesis. Two possibilities were considered for how L63 could be involved in this developmental process: (1) as a regulator of individual epithelial cell shape changes and (2) as a component of a communication system between epithelial cells. One mechanism by which L63 proteins could alter the changes in cell shape due to remodeling of the cytoskeleton is by phosphoregulating critical components of the cytoskeleton, such as actin. Alternatively, if L63 is part of a cell-cell communication system, the bent-leg phenotype of L63 mutants could be explained by a failure of the epithelial cells to communicate to each other their relative spatial positions, thereby altering the evagination process that occurs in the leg disc during pupal development. A role for L63 in the morphogenesis of the leg disc is further supported by recent genetic results that show that L63 mutations genetically interact with genes involved in leg morphogenesis. It should be emphasized that whatever role L63 plays in epithelial morphogenesis, it cannot be generalized to all discs given the finding that eyes composed exclusively of L63 mitotic clones show no morphological defects. The tissue distributions of L63 proteins just before and during metamorphosis also suggest a role for L63 in the cell death of the histolysing larval tissues. In this regard, it is noted that the larval midgut, which expresses little or no L63 protein in mid-third-instar larvae (18 h before pupariation), exhibits strong L63 expression at pupariation while in the midst of the process of cell death and histolysis. In this connection it is also noted that the gastric caeca, which undergo histolysis before the midgut, exhibit strong L63 expression earlier than does the midgut (Stowers, 2000).
There are three putative ecdysone-induced genes that have been identified in the L63 region: ImpE2, L63, and a gene with a 2.0-kb poly(A)+ transcript. L63 contains the DNA required for the formation of the 63E late puff. ImpE2 is a previously defined gene that exhibits a primary response to the late-larval ecdysone pulse in imaginal discs, but is not responsive to that pulse in salivary glands, and the gene with the 2-kb transcript was not analyzed further. A large set of cloned L63 cDNAs from three independent libraries provided the source for extensive restriction and selective sequence analyses that define the overlapping L63A, B, and C transcription units from which nine mRNAs are formed by alternative splicing. These mRNAs contain 10 common exons (Nos. 4-13), except that the A2 and B2 mRNAs lack exon 6 and are thereby distinguished from the A1 and B1 mRNAs, respectively. Otherwise, the mRNAs are distinguished by exons specific to each transcription unit, i.e., exons alpha1-alpha3, beta1-beta4, and gamma1-gamma3 for transcription units L63A, B, and C, respectively (Stowers, 2000).
In addition to the previously identified Drosophila cdc2 and cdc2c genes, four cdc2-related genes have been isolated with low stringency and polymerase chain reaction approaches. Sequence comparisons suggest that the four putative kinases represent the Drosophila homologs of vertebrate kinases Cdk4/6, cdk5, PCTAIRE, and PITSLRE. Although the similarity between human and Drosophila homologues is extensive in the case of cdk5, PCTAIRE, and PITSLRE kinases (78%, 58%, and 65% identity in the kinase domain), only limited conservation is observed for Drosophila Cdk4/6 (47% identity). However, like vertebrate cdk4 and cdk6, Drosophila cdk4/6 binds also to a D-type cyclin according to the results of two-hybrid experiments in yeast. Sequence conservation and expression patterns, therefore, suggest that all of these kinases perform important cellular functions (Sauer, 1996).
Determination of the nucleotide sequences of the nine L63 cDNAs and of genomic DNA sequences sufficient to define the exon boundaries within these cDNAs has shown that only five L63 protein isoforms are encoded by the nine L63 mRNAs. For example, the N-terminal block of the L63A1 isoform represents the amino acids encoded by the 39-terminal sequence of the overlapping a1 and a3 exons of the L63A1 and L63A3 mRNAs, respectively (a group of upstream exons coding for the 5' portion of L63A1 and L63A3 mRNAs). Because the splicing patterns downstream of alpha1 and alpha3 are identical for these two mRNAs (they are coded for by exons 4 through 13), they encode the same L63A1 protein. In contrast, the L63A2 mRNA, which differs from L63A1 only by the absence of exon 6, encodes the shorter L63A2 isoform (Stowers, 2000).
The L63B transcription unit encodes the three other protein isoforms (L63B1, 2, 3), one of which (B1) can also be encoded by the L63C transcription unit. Indeed, this isoform can be encoded by four mRNAs, two from each unit. The L63B1 isoform differs from L63A1 only in respect to the N-terminal amino acids that are encoded by the alpha1 or alpha3 exons for L63A1 and by the beta3, beta4, gamma1, or gamma3 exons for L63B1 (the beta and gamma exons are found positioned between the upstream alpha exons and the downstream common exons). The L63B2 and L63A2 isoforms also differ only by these two N-terminal amino acid sequences; in this case, however, each isoform appears to be encoded by a single mRNA. The L63B3 mRNA that encodes the remaining isoform, L63B3, is unique in that all of its codons derive from the common exons 4-13. Since there is no in-frame AUG in the beta1 exon of the L63B3 mRNA, translation is presumed to begin with the first in-frame AUG in exon 4 (Stowers, 2000).
Each of the five isoforms includes a 294-residue region (residues 199-492 of isoforms L63A1 and B1) that exhibits strong similarity (51% amino acid identity) to the CDC28 CDK of Saccharomyces cerevisiae. The L63 isoforms differ from CDC28 in having long N-terminal extensions of 176 to 197 residues. In this they are akin to certain human CDK proteins, called PCTAIRE-1, -2 and -3 (Meyerson, 1992). PCTAIRE has a 158-residue N-terminal extension followed by a CDK-homology region (residues 159-449) with 61% identity to that of L63. Of even more interest is a murine CDK protein called PFTAIRE that also has an N-terminal tail followed by a CDK region with a remarkable 70% identity to L63 (Lazzaro, 1997). The L63, PFTAIRE, and PCTAIRE-1 proteins are also similar in having C-terminal tails that extend beyond the C-terminus of CDC28. As distinguished from the N-terminal extensions, which exhibit little or no sequence similarity, the first 24 residues of the murine PFTAIRE and the human PCTAIRE-1 C-terminal tails exhibit, respectively, 54% and 42% identity to the equivalent L63 residues. By the above criteria, the L63 isoforms most closely resemble the murine PFTAIRE protein, which derives its name from the sequence of 7 amino acids found in the highly conserved PSTAIRE motif of the proven CDKs, such as the human CDK2 and the yeast CDC28 (residues 52-58). In the human CDK2, the PSTAIRE sequence is contained within the alpha1 helix, which undergoes a conformational change in the cyclin A-CDK2 complex that is critical to kinase activation. A single amino acid change from serine to cysteine, or to phenylalanine, in this signature sequence is responsible for the nomenclature of the human PCTAIRE-1 (residues 205-211) and murine PFTAIRE (residues 129-135) proteins, respectively. Here again, the L63 isoforms, which exhibit the PFTAIRE sequence motif (residues 245-251), are most similar to the mouse protein. Accordingly, the L63 isoforms could be called PFTAIRE proteins; indeed, this terminology was used for the L63A2 isoform, as defined by the sequence of a single cDNA (Sauer, 1996 and Stowers, 2000).
The N-terminal sequences of the L63 isoforms, while showing little or no similarity to other CDK proteins with N-terminal extensions, do exhibit certain interesting characteristics. One of these is the 9-residue polythreonine tract near the N-terminus of the L63A1 and A2 isoforms; it is somewhat curious that the quite different N-terminal sequences encoded by the alpha1 and beta3 exons each consist of 19 residues. A stretch of 19 residues encoded by exon 6 (residues 56-75) contains 13 histidines (68%), which includes a polyhistidine tract of 9 residues (Stowers, 2000).
date revised: 20 June 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.