Deletion mapping of sequences required for the ecdysone induction of the 63E late puff suggests that L63B transcription, and possibly that of L63C, is ecdysone responsive. Northern analyses were used to show that the temporal profile of L63B (Ecdysone-induced protein 63E) transcription mimics that of known ecdysone-responsive genes and that the L63A profile is more complex, while L63C transcription is too weak to be detected by these analyses. While sequence overlap exists between beta3 and the gamma1 exon of L63C, a probe specific for gamma1 sequences fails to detect any mRNAs on an identical Northern blot, even though it contains more L63C nucleotides than are present in the beta3 exon. The temporal profile of the L63B1/B2 mRNAs is like that of several ecdysone-responsive genes. The L63B1/B2 midembryonic expression is like that observed for the ecdysone-inducible early-late DHR3 gene and the EcR-A and EcR-B1 transcription units that encode respective ecdysone receptor isoforms. The embryonic ecdysone pulse presumed to induce these genes peaks at 10 h and is well suited to generating the L63B embryonic expression pattern. Similar arguments apply to the strong L63B1/B2 expression that extends from late third instar (L3) to the middle of pupal development (108-168 h) when DHR3 and EcR-A are also strongly expressed in conjunction with the late-larval, prepupal, and pupal ecdysone pulses that peak at 120, 130, and 150 h, respectively. The same temporal pattern of transcription is seen for the band of RNA for which the probe consists of sequences in the common 8-13 exons. This band was assigned to the 2.6-kb L63B3 mRNA because the match of lengths and pattern are unique (Stowers, 2000).
In contrast to L63B transcription, the observed temporal profile for the 3.8-kb L63A1 and L63A2 mRNAs indicates that, with one exception, the ecdysone pulses do not influence the abundance of these mRNAs in whole animals. That exception is an increase in transcript abundance during the 120- to 168-h interval that could result from the prepupal and/or pupal ecdysone pulses. It should, however, also be noted that this increase in the L63A mRNA might also result from the same regulatory mechanisms that cause a reappearance of the 63E puff during the middle of the prepupal period when the ecdysone levels are very low. Given that the 3.1-kb L63A3 mRNA is represented by the 3.0-kb band of the middle blot, then its expression is largely confined to embryogenesis. The large abundance of all of the L63A mRNAs during the first few hours of embryogenesis are probably of maternal origin, particularly because the abortion of transcription during the short nuclear division cycles of this period makes it highly unlikely that the long L63A primary transcript could be zygotically produced (Stowers, 2000).
Monoclonal anti-L63 antibodies divide into two classes, one class against epitopes in the N-terminal tail and the other class against epitopes in the C-terminal tail, with all epitopes present in all five L63 isoforms. Specificity criteria for L63 in the tissue staining are that the pattern obtained with an antibody from one class (A1-29; N-terminal epitope) is the same as that obtained with antibody from the other class (E3-53; C-terminal epitope). L63 is expressed throughout development and in many tissues. However, some tissues show different L63 expression levels at different developmental stages, especially at metamorphosis. Comparison of the larval midgut shows that the L63 protein increases from little or nothing in feeding mid-third-instar larvae (i.e., 18 h before pupariation) to high levels in white prepupae (i.e., at pupariation), implying that one or more of the L63 transcription units is activated in response to the late-larval pulse of ecdysone. The same developmental correlation of L63 with ecdysone titer is observed in ovaries that show no staining in mid-third-instar larvae; rather, by pupariation, they contain high levels of L63. The gastric caeca, which exhibit high levels of L63 at mid-third instar, show signs of programmed cell death by pupariation and disappear a few hours later. Another pattern of L63 expression is seen in both the larval salivary gland and the imaginal discs. In mid-third instar, both exhibit significant levels of L63 that increase severalfold by pupariation, as if L63 is induced in two phases. Interestingly, a similar developmental expression profile for L63 is observed in the ring gland, the site of ecdysone biosynthesis, with moderate expression levels at mid-third instar and high expression levels at white prepupae. L63 is primarily localized to the cytoplasm. A similar conclusion can also be drawn from close examination of the salivary gland staining. The observation that L63 protein is present in some tissues before the later larval ecdysone pulse is consistent with temporal patterns of L63A and total L63 mRNA expression and the predominantly premetamorphic lethal phase of L63 mutants (Stowers, 2000).
Evidence that L63 is an essential gene was obtained from the analysis of seven lethal EMS mutations in the 63E region. Molecular analysis of the 81 mutation shows that it is a mutant allele of the L63 gene, hereafter denoted L6381, consisting of an in-frame deletion of the DNA sequences encoding residues 226-241 within the conserved CDK region of the L63 isoforms. Western analyses of proteins in the L6381 mutants reveal the expected shift in mobility resulting from the deletion of 16 amino acids. One of these is the highly conserved lysine (K234) that forms an essential part of the ATP-binding catalytic triad in the human CDK2 protein. Hence, it is likely that the L6381 mutation is null for L63. The Df E1/Df GN50 transheterozygote is an even better candidate for an L63 null because Df E1 eliminates all three of the L63 transcription initiation sites along with adjacent, presumably regulatory DNA, while Df GN50 eliminates virtually all L63 coding sequences. The supposition that the Df E1/Df GN50 heterozygote is an L63 null is further confirmed by the inability to detect any L63 protein by Western analysis of this mutant genotype at metamorphosis. As expected, the Df E1/Df GN50 transheterozygote exhibits a lethal phase much like that of the L6381/Df GN50. Both mutant combinations show a heterogeneous lethal phase with the majority of mutant animals dying during larval development and a small percentage living to pupation with an occasional adult escaper. Further evidence that the lethality of these heterozygotes results from alteration of L63 is that the lethal L6381/DfGN50 and DfE1/DfGN50 heterozygotes can be rescued to adulthood by heat-shock-induced expression of the L63B1 isoform. This result provides strong evidence that defective L63 genes cause the above lethalities. No aberrant phenotypic consequences of the ectopic L63B1 expression were observed in these rescue experiments (Stowers, 2000).
The redundancy of the L63 isoform functions is indicated not only by the ability of a single isoform, L63B1, to rescue Df E1/Df GN50 embryos to adulthood, but also by the viability of deletion mutants in which the production of a particular isoform is eliminated. The most striking illustration of this redundancy is the finding that Df 1-11/Df GN50 animals are viable and can be perpetuated as a stock. Df 1-11 removes all of the exons specific to the L63A proteins (exons a1-3), as well as potential regulatory regions both upstream and downstream of the L63A transcription initiation site. Given that the Df GN50 is null for L63, the viability of Df1-11/Df GN50 animals indicates that the L63A1 and 2 proteins are not required when the L63B proteins are present. The reduction of L63 protein levels in Df1-11/Df GN50 late larvae, compared to the Canton-S wildtype, may result not only from the Df GN50 deletion, but also from the absence of L63A proteins due to the Df 1-11 deletion. If this is the case, then the protein band seen above the Df FF/Df GN50 main band, but not seen in Df 1-11/Df GN50, may represent one of the L63A isoforms (Stowers, 2000).
The Df FF and Df CC deficiencies eliminate the B1 exon and hence the L63B transcription initiation site. The observation that Df CC can be maintained indefinitely as a homozygote indicates that the L63B transcription initiation site and neighboring sequences are not required for viability. One cannot, however, conclude that the L63B1, B2, and B3 proteins are not required in the presence of the L63A1 and A2 proteins because the L63B1 protein is encoded by the L63C1 and 2 mRNAs as well as the L63B1 and B4 mRNAs. Furthermore, there is no obvious reason why an additional L63C mRNA encoding the L63B2 protein might not be made by the same exon 5/7 splice used for the L63A2 and L63B2 mRNAs. Indeed, the fact that such an L63C cDNA was not detected may well result from the low abundance of the L63C mRNAs in the wildtype (Stowers, 2000).
The rescue of Df E1/Df GN50 animals to adulthood provided an opportunity for carrying out a self-cross of w; HSL63/1; DfE1/DfGN50 animals. The observation that none of the embryos from this cross were observed to hatch demonstrates an embryonic requirement for the L63 protein. Moreover, this observation raises the possibility of a maternal effect because Df E1/Df GN50 mutants derived from females not maternally deficient for L63 die later in development. The viability of Df 1-11/Df GN50 animals (maternally deficient for L63A) allowed this possibility to be confirmed by showing that the direction of the cross Df 1-11/Df GN50 3 Df E1/TM6b affects the lethal phase of the L63 mutant genotype. When female Df 1-11/Df GN50 are crossed to male Df E1/TM6b, all Df E1/Df GN50 L63 mutant progeny die prior to the third-instar developmental stage, while if the reciprocal cross is performed a significant percentage of Df E1/Df GN50 mutant progeny survive to third instar or further, exhibiting a lethal phase similar to that of Df E1/Df GN50 animals from the cross Df E1/TM6b 3 Df GN50/TM6b. Notably, when L63 maternally deficient w; HSL63/1; DfE1/DfGN50 mothers are crossed to Df E1/TM6b males, TM6b progeny survive to adulthood. This result indicates that the embryonic requirement for L63 that is absent in the L63 maternal deficient females is zygotically rescuable by the wild-type L63 function present on the TM6b chromosome (Stowers, 2000).
Phenotypic analysis of rare L63 null mutant animals that survive past the third instar stage reveal developmental roles for L63. One such role is a general effect on the overall size of the animal. L63 mutant pupae range in size from approximately one-third to two-thirds that of wildtype under uncrowded conditions and take 2-3 days longer than their heterozygous siblings to reach pupariation. A second and more specific developmental role for L63 is an involvement in epithelial morphogenesis. This is evidenced by the bent-leg phenotype in rare escaper adult L63 mutant animals. While such L63 mutant animals do manage to eclose they almost immediately fall into the food and die soon thereafter. Both the small pupal size and the bent-leg phenotypes are rescued along with viability in the heat-shock-induced L63 cDNA rescue experiments. In contrast to the morphogenetic defect observed in L63 mutant legs, no morphogenetic defects are observed in L6381 mutant eyes composed exclusively of mitotic clones of L63 cells. Furthermore, analysis of the electrophysiological properties of L63 mutant eyes by electroretinogram reveals no defects in phototransduction or synaptic transmission (Stowers, 2000).
The transgene rescue technique was adapted to test for CDK function by attempting to rescue Df E1/Df GN50 animals with heat-shock-induced expression of mutant L63 proteins in place of wild-type L63B1. Thus, in these experiments a mutant version of the L63B1 cDNA was linked to the hsp70 promoter in the context of a P-element vector that was germ-line transformed to generate a mutant HS L63 chromosome. 32°C as well as 37°C heat shocks were used, thereby testing for rescue at different in vivo concentrations of the mutant L63 protein. Residues known to be important for cyclin binding, kinase activity, and phosphorylation in CDK proteins were chosen for site-directed mutagenesis. If these functional hallmarks of the CDK family contribute to the in vivo function of L63, it would be predicted that mutating these residues would adversely affect the ability of the altered L63 protein to rescue the Df E1/Df GN50 mutant. The conserved glycine (G243) and isoleucine (I249) residues of L63B1 correspond to human CDK2 residues known to be crucial for binding cyclin A. In the human CDK2/CYCA cocrystal structure, the glycine corresponding to the L63 G243 is adjacent to the cyclin-interacting PSTAIRE domain (PFTAIRE in L63). It is thought to be essential at that position because it is the only amino acid that will allow the two residues directly adjacent to it to form hydrogen bonds with the human cyclin A. The isoleucine residue in human CDK2 that corresponds to the L63 I249 is crucial because it fits tightly into a hydrophobic pocket of the human cyclin A. G243A L63 mutant protein fails to rescue the Df E1/Df GN50 animals exposed to either the 32°C or the 37°C heat shocks and the I249L mutant provides only partial rescue compared to the L63B1 wild-type (17% with 32°C heat shock and 68% with 37°C heat shock). These differences cannot be attributed to differences in the heat-shock response because the wild-type and mutant protein levels were similar for a given temperature. These results are taken as strongly supportive of the proposition that L63 protein interacts with cyclins (Stowers, 2000).
The glutamic acid residue at 251 in L63B1 (E251) corresponds to a highly conserved residue in both cyclin-dependent kinases and other, more distantly related, protein kinases. In the human CDK2, this residue assists the highly conserved lysine and aspartic acid residues (the other two members of the catalytic triad -- K234 and D344 in L63B1) in correctly positioning the gamma phosphate of ATP for nucleophilic attack by serine and threonine substrates. Mutation of the glutamic acid to glutamine results in complete loss of the rescue function at both temperatures -- a result in strong support of the minimal proposition that L63 is a protein kinase. The serine residue S359 in L63B1 corresponds to a conserved threonine phosphorylation site in human CDK2 (T160) and in other CDKs, including the T169 site in the yeast CDC28. Phosphorylation of this residue is critical for full CDK activity. The L63B1 mutant in which this serine is replaced by alanine (S359A) cannot rescue Df E1/Df GN50 whether induced at 32°C or 37°C (Stowers, 2000).
An attempt was made to see if the N-terminal tail plays an essential role in L63 function by constructing a mutant, delta202 L63, in which the N-terminal 202 residues of L63B1 were eliminated. No rescue by delta202 L63 was observed when the heat shocks were carried out at 32°C, but 31% rescue occurs with 37°C heat shocks. This result cannot, however, be directly compared with those of the other L63 rescue constructs because the amount of delta202 L63 stably produced by the 37°C heat shocks is only 15% of that obtained for the other L63 constructs. What can be said about the nonconserved N-terminal tail is that it is not necessary for L63 function, although it may enhance that function or protect L63 from denaturation and/or destruction. An ancillary experiment was carried out in which it was asked whether either of the bona fide Drosophila CDKs (cdc2 and cdc2c) could rescue the Df E1/Df GN50 null. The answer was a clear no, even though both proteins were produced at high levels by the 37°C heat shocks (Stowers, 2000).
Besset, V., Rhee, K. and Wolgemuth, D. J. (1998). The identification and characterization of expression of Pftaire-1, a novel Cdk family member, suggest its function in the mouse testis and nervous system. Mol. Reprod. Dev. 50(1): 18-29.
Carrel, L., et al. (1996). X inactivation analysis and DNA methylation studies of the ubiquitin activating enzyme E1 and PCTAIRE-1 genes in human and mouse. Hum. Mol. Genet. 5(3): 391-401.
Charrasse, S., et al. (1999). PCTAIRE-1: characterization, subcellular distribution, and cell cycle-dependent kinase activity. Cell Growth Differ. 10(9): 611-20.
Gao, C. Y., et al. (1996). Expression of alternatively spliced PCTAIRE-1 mRNA in PC12 cells and neonatal rat brain. Gene 176(1-2): 243-7.
Hirose, T., et al. (1997). PCTAIRE 2, a Cdc2-related serine/threonine kinase, is predominantly expressed in terminally differentiated neurons. Eur. J. Biochem. 249(2): 481-8.
Hirose, T., et al. (2000). Identification of tudor repeat associator with PCTAIRE 2 (Trap). A novel protein that interacts with the N-terminal domain of PCTAIRE 2 in rat brain. Eur. J. Biochem. 267(7): 2113-21.
Lazzaro, M. A., Albert, P. R. and Julien, J. P. (1997). A novel cdc2-related protein kinase expressed in the nervous system. J. Neurochem. 69(1): 348-64.
Le Bouffant, F., et al. (1998). Characterization of brain PCTAIRE-1 kinase immunoreactivity and its interactions with p11 and 14-3-3 proteins. Eur. J. Biochem. 257(1): 112-20.
Le Bouffant, F., et al. (2000). Multiple subcellular localizations of PCTAIRE-1 in brain. Mol. Cell. Neurosci. 16(4): 388-95. 11085876
Meyerson, M., et al. (1992). A family of human cdc2-related protein kinases. EMBO J. 11(8): 2909-17.
Michaelis, C. and Weeks, G. (1993). The isolation from a unicellular organism, Dictyostelium discoideum, of a highly-related cdc2 gene with characteristics of the PCTAIRE subfamily. Biochim. Biophys. Acta. 1179(2): 117-24.
Okuda, T., et al. (1994). Cloning of genomic loci and chromosomal localization of the human PCTAIRE-1 and -3 protein kinase genes. Genomics 21(1): 217-21.
Rhee, K. and Wolgemuth, D. J. (1995). Cdk family genes are expressed not only in dividing but also in terminally differentiated mouse germ cells, suggesting their possible function during both cell division and differentiation. Dev. Dyn. 204(4): 406-20.
Sauer, K., et al. (1996). Novel members of the cdc2-related kinase family in Drosophila: cdk4/6, cdk5, PFTAIRE, and PITSLRE kinase. Mol. Biol. Cell. 7(11): 1759-69.
Sladeczek, F., et al. (1997). The Cdk-like protein PCTAIRE-1 from mouse brain associates with p11 and 14-3-3 proteins. Mol. Gen. Genet. 254(5): 571-7.
Stowers, R. S., et al. (2000). The L63 gene is necessary for the ecdysone-induced 63E late puff and encodes CDK proteins required for Drosophila development. Dev. Biol. 221(1): 23-40.
date revised: 20 June 2001
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.