Little is known about how patterns of cell proliferation and arrest are generated during development, a time when tight regulation of the cell cycle is necessary. In this study, the mechanism by which the developmental signaling molecule Wingless generates G1 arrest in the presumptive Drosophila wing margin is examined in detail. Wg signaling promotes activity of the Drosophila retinoblastoma family (Rbf) protein, which is required for G1 arrest in the presumptive wing margin. Wg promotes Rbf function by repressing expression of the G1-S regulator Drosophila myc (dmyc). Ectopic expression of dMyc induces expression of Cyclin E, Cyclin D, and Cdk4, which can inhibit Rbf and promote G1-S progression. Thus, G1 arrest in the presumptive wing margin depends on the presence of Rbf, which is maintained by the ability of Wg signaling to repress dmyc expression in these cells. In addition to advancing the understanding of how patterned cell-cycle arrest is generated by the Wg signaling molecule during development, this study indicates that components of the Rbf/E2f pathway are targets of dMyc in Drosophila. Although Rbf/E2f pathway components mediate the ability of dMyc to promote G1 progression, dMyc appears to regulate growth independently of the RBF/E2f pathway (Duman-Scheel, 2004).
The results indicate why exclusion of dMyc from the ZNC is necessary for Rbf activity. Overexpression of dMyc leads to high levels of Cyclin E, Cyclin D, and Cdk4 transcripts. dMyc also regulates Cyclin E posttranscriptionally in Drosophila. G1-S Cyclins/Cdks function to phosphorylate and inhibit Rbf, suggesting that dMyc blocks Rbf activity through activation of G1-S Cyclins/Cdks. Thus, inhibition of dMyc by Wg helps to ensure that G1-S Cyclins/Cdks do not activate S phase. This idea is supported by the results that indicate that only a combination of both Dap and constitutively active Rbf (that cannot be regulated by Cyclins/Cdks) can restore G1 arrest when Wg signaling is blocked or when dMyc is expressed. These data suggest that either Cyclin D or Cyclin E activity can mediate the ability of dMyc to promote S phase in the ZNC. Coexpressing Dap alone with dMyc, which would block only Cyclin E/Cdk2 activity, does not restore G1 arrest. Furthermore, overexpression of dMyc in a cdk4 mutant background still results in ectopic S phases, suggesting that Cyclin E/Cdk2 also are sufficient to mediate dMyc's ability to promote G1 progression. Thus, either Cyclin D/Cdk4 or Cyclin E/Cdk2 is sufficient to mediate the ability of dMyc to promote G1 progression. The ability of Wg to inhibit dMyc expression is thus critical for RBF activation and G1 arrest in the ZNC. Still, it is possible that Wg promotes G1 arrest through other mechanisms that have not yet been uncovered. The observation that overexpression of a dominant-negative form of dTCF (dTCFDeltaN) with C96>Gal4 can promote S phase, even in a dmyc mutant background, supports this idea (Duman-Scheel, 2004).
It is likely that dMyc/dMax directly up-regulate transcription of Cyclin D and cdk4 in Drosophila. Myc/Max heterodimers regulate transcription by binding to various consensus sequences, such as the E box. Previous studies indicated that cMyc induces Cyclin D2 expression in mice by binding to two consensus E boxes in the Cyclin D2 promoter. cdk4 also was identified as a transcriptional target of c-Myc. Furthermore, it has been suggested that cdk4 is a transcriptional target of dMyc and Cyclin D is a transcriptional target of dMax. Although future studies should analyze the Drosophila Cyclin D and Cdk 4 regulatory regions in more detail, these results suggest that the observed ability of dMyc to induce Cyclin D and Cdk4 expression in the ZNC most likely occurs through transcriptional regulation of these proteins by dMyc/dMax. In contrast, Cyclin E was not identified as a target of dMyc or dMax. It is more likely that the ability of dMyc to induce growth in the wing indirectly leads to increased Cyclin E transcript levels (Duman-Scheel, 2004).
Terminal differentiation is often coupled with permanent exit from the cell cycle, yet it is unclear how cell proliferation is blocked in differentiated tissues. The process of cell cycle exit was examined in Drosophila wings and eyes; cell cycle exit can be prevented or even reversed in terminally differentiating cells by the simultaneous activation of E2F1 and either Cyclin E/Cdk2 or Cyclin D/Cdk4. Enforcing both E2F and Cyclin/Cdk activities is required to bypass exit because feedback between E2F and Cyclin E/Cdk2 is inhibited after cells differentiate, ensuring that cell cycle exit is robust. In some differentiating cell types (e.g., neurons), known inhibitors including the retinoblastoma homolog Rbf and the p27 homolog Dacapo contribute to parallel repression of E2F and Cyclin E/Cdk2. In other cell types, however (e.g., wing epithelial cells), unknown mechanisms inhibit E2F and Cyclin/Cdk activity in parallel to enforce permanent cell cycle exit upon terminal differentiation (Buttitta, 2007).
Current models for cell cycle exit invoke repression of Cyclin/Cdk activity by CKIs or repression of E2F-mediated transcription by RBs as the proximal mechanisms by which cell cycle progression is arrested. Since these models include the potential for positive feedback between E2F and CycE/Cdk2, they predict that the induction of either E2F or a G1 Cyclin/Cdk complex should be sufficient to maintain the activity of the other and thereby sustain the proliferative state. However, in differentiating Drosophila tissues, both E2F and G1 Cyclin/Cdk activities had to be simultaneously upregulated to bypass or reverse cell cycle exit. An explanation for this resides in two observations. First, the ability of Cyclin/Cdk activity to promote E2F-dependent transcription is lost or reduced in the wing and eye after terminal differentiation. Second, increased E2F cannot sustain functional levels of CycE/Cdk2 activity after terminal differentiation, despite an increase in cycE and cdk2 mRNA to levels higher than those observed in proliferative-stage wings. Thus, crosstalk between E2F and Cyclin/Cdk activity appears to be limited, in both directions, as a consequence of differentiation (Buttitta, 2007).
How are these two regulatory interactions altered? One possibility is that Rbf2- or E2F2-dependent repression prevents ectopic Cyclin/Cdk activity from promoting E2F-dependent transcription after prolonged exit. While mRNA expression data and the existing genetic data on E2F2 and Rbf2 do not support this possibility, the roles of Rbf2 or E2F2 have not been tested in the presence of continued Cyclin/Cdk activity. Therefore, transcriptional repression of E2F targets by Rbf2 or E2F2 remains an important issue to address in future experiments (Buttitta, 2007).
More enigmatic is the inability of the ectopic CycE/Cdk2 provided by overexpressed E2F to promote cell cycle progression. One plausible explanation for this is that novel inhibitors of CycE are expressed with the onset of differentiation, and that these raise the threshold of Cyclin/Cdk activity required to promote cell cycle progression. Such inhibitors might make the critical substrates of CycE/Cdk2, which reside on chromatin in DNA-replication and -transcription initiation complexes, less accessible or otherwise recalcitrant to activation. The notion of an increased Cdk threshold is consistent with the observation that the >10-fold increase in CycE/Cdk2 provided by direct overexpression of the kinase bypassed cell cycle exit in conjunction with E2F, while the ~4-fold increase provided indirectly by ectopic E2F is insufficient to drive the cell cycle. Although a >10-fold increase in Cdk activity as applied in these experiments is far above the normal physiological range, such dramatic deregulation of cell cycle genes may be physiologically relevant to cancers, in which gene expression can be greatly amplified (Buttitta, 2007).
Recent studies of cycle exit in larval Drosophila eyes have concluded that Rbf1 and Dap are required to inhibit E2F and CycE/Cdk2 in differentiating photoreceptors. Other studies document the roles of Ago/Fbw7 and components of the Hippo/Warts-signaling pathway in downregulating CycE for cell cycle exit in nonneural cells in the eye. Although the data are consistent with these studies in the eye, Ago and the Hippo/Warts pathway are dispensable for cell cycle exit in the wing. Furthermore, deletion of Rbf1 did not prevent cell cycle exit in the epithelial wing, even when high levels of CycE/Cdk2 were provided. Conversely, deletion of Dap was not sufficient to keep wing cells cycling, even when excessive E2F activity was provided. These observations suggest that unknown inhibitors of E2F and Cyclin/Cdk activity mediate cell cycle exit in specific contexts, such as the wing (Buttitta, 2007).
In attempts to identify upstream factors regulating cell cycle exit, a variety of growth and patterning signals were manipulated in the pupal wing and eye, and their effects on cell cycle exit were examined. Surprisingly, signals that act as potent inducers of proliferation in wings and eyes at earlier stages did not prevent or even delay cell cycle exit upon terminal differentiation. Thus, an important focus for future studies will be the nature of the signals upstream of E2F and CycE that mediate cell cycle exit. These could be novel signals, or combinations of known signals delivered in unappreciated ways (Buttitta, 2007).
How general is double assurance? Studies of cell cycle exit in mammals do not offer a consistent answer to this question. S phase re-entry can be achieved in differentiated cells by activating E2F, CycE/Cdk2, or CycD/Cdk4 alone, but this does not lead to cell division or continued proliferation. Several studies with mammalian cells in vivo have shown that neither increased E2F nor Cyclin/Cdk activity alone is sufficient to fully reverse differentiation-associated quiescence, consistent with the double-assurance model propose in this study. Also consistent with this model is the ability of proteins from DNA tumor viruses, such as adenovirus E1A, SV40 LargeT, and HPV E6 and E7, to fully reverse differentiation-associated cell cycle exit in many cell types. These viral onco-proteins stimulate cell cycle progression by targeting multiple cell cycle factors, which ultimately increase both E2F and G1 Cyclin/Cdk activities simultaneously. For example, LargeT and E1A inhibit both RBs and CKIs, such as p21Cip1 and p27Kip1 (Buttitta, 2007).
There are some instances, however, in which differentiation-associated cell cycle exit has been bypassed, not just delayed, by the deletion of CKIs or RBs. In one such case, p19Ink4d and p27Kip1 were knocked out in the mouse brain, and ectopic mitoses were documented in neuronal cells weeks after they normally become quiescent. Similar results have been obtained with hair and support cells in the mouse inner ear, where deletion of p19Ink4d, p27Kip1, or pRB can bypass developmentally programmed cell cycle exit. In light of these findings, it is interesting to speculate that certain differentiated tissues may retain some ability to repair or regenerate by maintaining the capacity for positive feedback between E2F and CycE/Cdk2 activity. Inner-ear hair cells may be such an example, since in many vertebrates they are capable of regeneration, although this ability has been lost in mammals. Although the mammalian brain has a very limited capacity for regeneration, the cell cycle can be reactivated in the brains of other vertebrates, such as fish, in response to injury. Thus, the retention of crosstalk between E2F and Cyclin/Cdk activities in the evolutionary descendents of regeneration-competent cells might explain some of the tissue-specific sensitivities to loss of CKIs or RBs observed in mammals (Buttitta, 2007).
The characteristic pRb-binding motif LXCXL is present in Drosophila Cyclin D. Co-immunoprecipitation experiments confirmed that Drosophila Cyclin D associates with Cdk4 in vivo. Immunoblotting with a monoclonal antibody reveals the presence of Cyclin D in anti-myc immunoprecipitates isolated from extracts of Cdk4-myc-expressing embryos. These results indicate therefore that Cyclin D binds specifically to Cdk4. The CIP/KIP-type inhibitor Dacapo does not associate with Cdk4 (Meyer, 2000).
D-type cyclin complexes are thought to be the major Rb kinase in mammalian cells. Therefore it was of interest to compare Rb kinase activity in wild-type and Cdk4mutant embryos. Unfortunately, the antibodies used failed to precipitate Rb kinase activity even from wild-type embryo extracts. Moreover, in contrast to vertebrate Rb, Drosophila RBF does not change its apparent electrophoretic mobility upon phosphorylation. As also observed with mammalian Rb, Drosophila RBF could not be focused on two-dimensional gels. Thus, using da-GAL4 mouse Rb was expressed in Drosophila embryos from a UAS transgene (UAS-mRb). Immunoblotting with an antibody against Rb clearly reveals two forms with different electrophoretic mobility. Phosphatase treatment converts the lower into the higher mobility form. Conversely, co-expression of UAS-Cdk4 and UAS-Cyclin D results in a relative increase in the lower mobility form. These observations suggest that Cyclin D-Cdk4 might in principle function as Rb kinase. However, no decrease in the abundance of the low mobility form could be detected in extracts from embryos lacking both maternal and zygotic Cdk4 function. It appears therefore that kinases other than Cyclin D-Cdk4 can phosphorylate mRb. In fact, experiments involving expression of either UAS-Cyclin E or UAS-dacapo suggest that Cyclin E-Cdk2 is a major Rb kinase in Drosophila embryos. UAS-Cyclin E results in a strong enrichment of the lower mobility form. In contrast, UAS-dacapo results in a severe reduction of the lower mobility form (Meyer, 2000).
Inactivation of Cyclin E-Cdk2 is essential for a timely arrest of the epidermal cell proliferation program during Drosophila embryogenesis. E-type cyclin-cdk complexes are thought to be activated by D-types titrating away inhibitors and inducing cyclin E transcription by activating E2F transcription factors via Rb phosphorylation. Therefore, a study was performed to see whether the developmentally controlled inactivation of Cyclin E-Cdk2, required for the epidermal cell proliferation arrest, occurs as a consequence of Cyclin D-Cdk4 inactivation. However, preventing Cyclin D-Cdk4 inactivation by overexpression has a minimal effect on Cyclin E expression and does not interfere with the initial G1 arrest, while it readily induces the E2F target RnrS in arresting epidermal cells. Prolonged Cyclin D-Cdk4 overexpression eventually interferes with maintenance of quiescence in some cells. Moreover, in Cdk4 mutant embryos, some RnrS expression is still induced by Cyclin E overexpression, and endogenous Cyclin E expression as well as cell cycle progression is not affected, except for late aspects of the endoreduplication program. These findings argue against the proposed necessity of complete Rb inactivation by sequential phosphorylation by D- and E-type cyclin-cdk complexes. They demonstrate that Cyclin D-Cdk4 does not function as the master regulator of the embryonic cell proliferation program (Meyer, 2002).
In conclusion, while Cyclin D-Cdk4 overexpression interferes in some cells with maintenance of quiescence, it does not preclude an initial timely entry into proliferative quiescence. Moreover, the effects of Cyclin D-Cdk4 and Cyclin E overexpression in Cyclin E and Cdk4 mutant embryos, respectively, cannot be easily reconciled with an obligatory sequential action of D- and E-type cyclin-cdk complexes, which has been postulated for some classes of E2F target genes based on evidence obtained in vitro and in mammalian cell culture systems. Cyclin D-Cdk4 plays a minor role in the control of Cyclin E-Cdk2 activity and cell cycle progression during the embryonic cell division cycles. These cycles are characterized by a relaxed coupling to cell growth because of the abundant maternal stores that are stockpiled in the egg. However, Cyclin D-Cdk4 is partially required for progression through endoreduplication cycles late in embryogenesis, which might already be triggered by nutrients extracted from the yolk by the newly formed gut, similar to the larval endoreduplication cycles, which are clearly nutrient dependent. Moreover, consistent with the proposal that Cyclin D-Cdk4 complexes are primarily involved in the regulation of growth rates, oogenesis and imaginal development, which are tightly coupled to growth, are most severely affected in Cdk4 mutants (Meyer, 2002).
The JAK/STAT signal transduction pathway regulates many developmental processes in Drosophila. However, the functional mechanism of this pathway is poorly understood. The Drosophila cyclin-dependent kinase 4 (Cdk4) exhibits embryonic mutant phenotypes identical to those in the Hopscotch/JAK kinase and stat92E/STAT mutations. Specific genetic interactions between Cdk4 and hop mutations suggest that Cdk4 functions downstream of the HOP tyrosine kinase. Cyclin D-Cdk4 (as well as Cyclin E-Cdk2) binds and regulates STAT92E protein stability. STAT92E regulates gene expression for various biological processes, including the endocycle S phase. These data suggest that Cyclin D-Cdk4 and Cyclin E-Cdk2 play more versatile roles in Drosophila development (Chen, 2003).
In a large screen for autosomal P element-induced zygotic lethal mutations associated with specific maternal effect lethal phenotypes, a mutation, l(2)sh0671, located at 53C, was identified that showed a maternal effect segmentation phenotype. The phenotype is similar to the effect of loss of hop and stat92E gene activity during oogenesis. The P element, l(2)sh0671, was inserted into the second intron of the Cdk4 gene before the ATG translation initiation code (Chen, 2003).
The Cdk43 allele is an intragenic deletion of the Cdk4 gene, which eliminates the essential kinase domains. Cdk43 germline clone (GLC) phenotypes were examined; 70% of Cdk43 GLC embryos have strong 'hop-like' segmentation defects. As is the case with hop and stat92E embryos, the paternally rescued Cdk43 embryos show a consistent deletion of the fifth abdominal segment and the posterior mid-ventral portion of the fourth abdominal segment. Cdk43 null embryos have other defects, including reduction of the second thoracic and eighth abdominal denticle bands, fusion of the sixth and seventh bands, and head defects. The other 30% of embryos display other defects, such as poor cuticle development and dorsal open, that differ from alterations in components of the HOP/STAT92E pathway, suggesting that Cdk4 also functions in developmental processes other than the HOP/STAT92E pathway. The maternal effect phenotype for Cdk4 mutants is partially paternally rescuable. Expression of a Cdk4 cDNA encoding the full-length Cdk4 protein under the control of a heat-shock promoter in transgenic flies fully rescued the segmentation defects of Cdk43 GLC embryos at 25°C. Cdk4D175N is a dominant-negative form of Cdk4; the D175N mutation affects an aspartate residue that is required for the phosphotransfer reaction. Expression of the dominant-negative form of the cDNA under a maternal gene promoter generated the 'hop-like' segmentation defects. These data suggest that the observed segmentation defects are caused by disruption of the Cdk4 gene (Chen, 2003).
The similarity of the Cdk4 mutant phenotype to that of hop and stat92E suggests that these latter genes are involved in the same developmental process. A prediction of this hypothesis is that mutations in Cdk4 would affect the expression of segmentation genes in the same manner as hop and stat92E. The removal of either hop or stat92E activity is known to result in the stripe-specific loss of expression of several pair-rule genes. The enhancer elements responsible for control of the third stripe of eve expression have been mapped to a 500 bp element upstream of the eve transcriptional start site. A reporter gene construct containing a 5.2 kb eve promoter element driving lacZ shows expression of lacZ in eve stripes 2, 3, and 7. Removal of maternal activity of either hop or stat92E results in the loss of the third stripe from the reporter construct. Similarly, removal of maternal activity of Cdk4 also causes the specific loss of the third stripe, without affecting the second or seventh stripes (Chen, 2003).
The HOP/STAT92E pathway regulates tracheal formation through regulating trachealess (trh) gene expression in the embryo. It was reasoned that Cdk4 might also regulate tracheal formation. Tracheal formation was examined in wild-type, hop, and Cdk4 embryos by using an antibody [(mAb)2A12] that stains tracheal branches and trunks. In paternally rescued hop and cdk4 embryos, a similar defective tracheal system was formed that generally had several disruptions in the main trunk and several branches. These data suggest that Cdk4 regulates tracheal formation in a manner similar to the HOP/STAT92E signal transduction pathway (Chen, 2003).
To determine whether hop and Cdk4 genetically interact, a test was performed to see whether a reduction in the amount of maternal Cdk4 gene activity could enhance the maternal effect associated with a partial loss of function in the hop mutation. Embryos that are derived from mothers that carry GLCs of the hopmsv1 hypomorphic allele show weak segmentation defects, and many of them hatch. However, when these embryos are derived from females that also carry a single copy of Cdk43, they exhibit segmentation defects that are similar to embryos derived from females that lack all maternal hop activity, and none of them hatch. This result suggests that hop and Cdk4 act in concert to regulate embryonic segmentation (Chen, 2003).
Whether Cdk4 operates upstream or downstream of HOP was examined by testing whether the effect of a hyperactive hop allele could be negated by a reduction in the amount of Cdk4 gene activity. If Cdk4 is required for transducing the HOP signal, then reduction of Cdk4 should suppress a hop gain-of-function phenotype. The dominant temperature-sensitive hop allele, hopTum-l, was used for this experiment. When grown above 29°C, flies heterozygous for hopTum-l have reduced viability and the emerging adults develop melanotic tumors. Viability and formation of melanotic tumors at 29°C were compared in females heterozygous for hopTum-l and Cdk4 with females heterozygous only for hopTum-l. An improved survival rate was obtained by removing a single copy of Cdk4 in hopTum-l heterozygous females. However, the formation of melanotic tumors is affected less by removing a single copy of Cdk4 in hopTum-l heterozygous females (Chen, 2003).
To further examine the function of Cdk4 in the HOP/STAT92E signal transduction pathway, the genetic interactions of Cdk4 with hop and stat92E were tested in embryos. The hop (hopC111) and stat92E (stat92E6346) null embryos show a consistent deletion of the fifth abdominal segment and the posterior mid-ventral portion of the fourth abdominal segment, and none of them hatch. When HS-Cdk4 is ubiquitously expressed in hopC111 embryos, most embryos have complete fourth and fifth abdominal segments, and many of them hatch. Ubiquitous expression of Cdk4 has no effect in stat92E mutant embryos (Chen, 2003).
In mammals and Drosophila, Cdk4 forms a protein complex that regulates the cell cycle progression. The Cyclin D and Cdk4 complex (CycD-Cdk4) phosphorylates and releases RB from RB/E2F; free E2F then activates gene expression, including Cyclin E (CycE). Cyclin E and Cdk2 form a complex (CycE-Cdk2) and regulate the cell cycle at the G1-S transition point. To further examine relations between the HOP/STAT92E signal transduction pathway and cell cycle regulation, the genetic interaction of hop with CycE was tested. Like HS-Cdk4, HS-CycE rescues hopC111 embryo segmentation defects but has no effect on stat92E mutant embryos (Chen, 2003).
The viability and formation of melanotic tumors at 29°C were compared in females heterozygous for hopTum-l and CycE with females heterozygous only for hopTum-l. An improved survival rate was observed by removing a single copy of CycE in hopTum-l heterozygous females. As in the case of Cdk4, the formation of melanotic tumors is less affected by removing a single copy of CycE in hopTum-l heterozygous females. These results suggest that CycD-Cdk4 and CycE-Cdk2 complexes are members of the HOP/STAT92E signal transduction pathway and function downstream of the HOP tyrosine kinase and either upstream of or parallel to the STAT92E transcription factor (Chen, 2003).
Thus Cdk4 functions in the HOP/STAT92E pathway and regulates embryonic segmentation, tracheal formation, eye development, and melanotic tumor formation. Specific genetic interactions between Cdk4 and hop or stat92E mutations suggest that Cdk4 functions upstream of STAT and parallel to or downstream of the HOP tyrosine kinase. Furthermore, CycD-Cdk4 and CycE-Cdk2 bind and regulate STAT92E protein stability. These data demonstrate that, besides their role in regulating the cell cycle, CycD-Cdk4 and CycE-Cdk2 have a role in regulating cell fate determination and proliferation via STAT signaling (Chen, 2003).
STAT92E binds directly to the promoter of pair-rule genes and regulates their expression for segmentation. This occurs during the first 13 embryonic cell cycles, which are nearly synchronous and lack G1 and G2 gap phases. Obviously, the function of CycD-Cdk4 and CycE-Cdk2 is not to regulate the cell cycle during this period. The CycD-Cdk4 and CycE-Cdk2 complexes may regulate pair-rule gene expression through stabilizing STAT92E protein and increasing its transcription activity (Chen, 2003).
Expression of a kinase-impaired mutant form of Cdk4 (Cdk4D175N) using a maternal driver generates the 'hop-like' segmentation defects, indicating that the segmentation phenotype requires Cdk4 kinase activity. There is also a potential consensus sequence (SPVKR) of Cdk4/Cdk2 phosphorylation in the C terminus of STAT92E; however, mutation of S to A in the sequence does not significantly affect STAT92E activity in regulating gene expression, and attempts to detect STAT92E phosphorylation by Cdk4 have not been successful. CycD-Cdk4 and CycE-Cdk2 may simply bind and enrich nuclear STAT92E protein for its transcriptional activity. Regardless, the biochemical findings suggest that Cyclin-Cdk activity could affect STATs directly. STAT92E may have many transcriptional targets in the Drosophila genome, including genes involved in cell cycle regulation (such as RNrS). CycD-Cdk4 and CycE-Cdk2 can execute their versatile roles through regulating STAT92E protein. Cell cycle regulation may be one of their various functions (Chen, 2003).
A fraction of Cdk4 mutant embryos displays other defects that differ from alterations in components of mutations of the HOP/STAT92E pathway, suggesting that CycD-Cdk4 and CycE-Cdk2 have other targets besides STAT92E in Drosophila. RBF/E2F1 has been shown to be another target of the CycE-Cdk2 complex (Chen, 2003).
Excess HOP/STAT92E signaling induces cell overproliferation in the eye. Excess CycD-Cdk4 activity blocks differentiation and induces overgrowth in the eye. This study shows that excess HOP/STAT92E signaling can synergize with both CycD-Cdk4 and CycE-Cdk2 in melanotic tumors, but specifically synergizes with CycD-Cdk4, not CycE-Cdk2, to promote formation of an enlarged eye with extra ommatidia. Specifically, overexpression of either CycD-Cdk4 or CycE under the GMR-Gal4 driver significantly increased the fraction of S and G2 phase cells posterior to the MF at the expense of the G1-arrested cells. Both cyclins perturbed the normal program of cell cycle exit at differentiation. However, only CycD-Cdk4 appears to affect both cell cycle progression and cellular growth, whereas CycE affected only cell cycle progression (Chen, 2003).
The Drosophila cyclin-dependent protein kinase complex Cyclin D/Cdk4 induces cell growth (accumulation of mass) as well as proliferation (cell cycle progression). To understand how CycD/Cdk4 promotes growth, a screen was performed for modifiers of CycD/Cdk4-driven overgrowth in the eye. Loss-of-function mutations in Hif-1 prolyl hydroxylase (Hph: CG1114), an enzyme involved in the cellular response to hypoxic stress, dominantly suppress the growth but not the proliferation function of CycD/Cdk4. hph mutant cells are defective for growth, and, remarkably, ectopic expression of Hph is sufficient to increase cellular growth. Epistasis analysis places Hph downstream of CycD/Cdk4. Overexpressed CycD/Cdk4 causes an increase in Hph protein in tissues where Hph induces growth, suggesting a mechanism whereby Hph levels are regulated posttranscriptionally in response to CycD/Cdk4. These data suggest that Hph, in addition to its function in hypoxic response, is a regulator of cellular growth and that it is a key mediator for CycD/Cdk4. Nevertheless, observations weigh against an important role for Hif-1 in Hph-driven growth, suggesting a potential Hif-1-independent function of Hph (Frei, 2004).
The synthetic promoter element Glass multiple reporter (GMR) is expressed in the eye imaginal disc posterior to the morphogenetic furrow, where only cells in the second mitotic wave undergo one synchronized cell division. GMR can be used to drive expression of the yeast transcription factor Gal4. Gal4 can be used to direct transcription of transgenes placed next to the UAS binding site of Gal4. Therefore, UAS transgenes driven by Glass-activated GMR-Gal4 are expressed predominantly in postmitotic cells. Under these circumstances, expression of CycD/Cdk4 leads to an enlargement of the adult eye, bigger ommatidia and bristles, and a general rough appearance. Although some ommatidia have additional cells, the main cause of the enlargement is an increase in cell size leading to 61% larger ommatidia. In order to identify loss-of-function mutants that modify this phenotype, CycD/Cdk4 was expressed in a deficiency collection background and screened under the light microscope for modifiers. Out of 162 deficiencies that cover 60%-70% of the genome, four deficiencies were isolated that dominantly suppress CycD/Cdk4. Df(3R)6-7, which deletes polytene segments 82D3/8-F3/6, led to a decrease in the enlargement of the eye and bristle size. Most strikingly, under these conditions, CycD/Cdk4 led to an increase in ommatidia size of only 17. Subsequently, partial overlapping deficiencies were tested: Df(3R)3-4 showed the same suppression phenotype but not a third deficiency, Df(3R)110, demonstrating that the gene of interest is between 82F3 and F7. All available mutants in this region were tested, and two lethal P element insertions were identified that showed an identical suppression phenotype compared to the deficiencies: l(3)02255 is inserted 104 bp, and l(3)S030304 is inserted 1111 bp upstream of the putative transcription start site of Hph/dmHph/CG1114 (Frei, 2004).
Whether l(3)02255 and l(3)S030304 are alleles of hph and whether their loss of function causes the suppression of CycD/Cdk4 were addressed. Northern blot experiments revealed that both heterozygous mutants l(3)02255 and l(3)S030304 have reduced expression of hph compared to the endogenous levels and that l(3)02255/l(3)S030304 transheterozygotes lack all detectable hph expression. In addition to these loss-of-function mutants, the EP3200 line has an EP insertion 382 bp upstream of Hph. Expression of hph using this EP element, by the hs-Flp Act>CD2>Gal4 system, led to a weak expression of hph in the absence of a heat shock, due to leakage of the system. A further increase in hph expression occured upon heat shock (Frei, 2004).
A full-length Hph cDNA was cloned under the control of a UAS promoter and injected into flies. These Hph transgenes suppress at least partially the cell growth phenotype of homozygous hph mutants and completely suppress the reduced viability of l(3)02255/l(3)S030304 transheterozygotes. Furthermore, when UAS-Hph transgenes were coexpressed with CycD/Cdk4 using the GMR-Gal4 driver, the dominant suppression of CycD/Cdk4-driven overgrowth by l(3)02255 was inhibited and the overgrowth phenotype was restored. It is concluded that l(3)02255 and l(3)S030304 are alleles of Hph and that the reduction of Hph can suppress CycD/Cdk4-induced overgrowth (Frei, 2004).
Whether growth or proliferation would be suppressed by hph in the eye imaginal disc was tested. GFP was expressed either alone or together with CycD/Cdk4 in wild-type, hph02255/+, or Df(3R)3-4/+ backgrounds using the GMR-Gal4 driver. Imaginal eye discs from wandering third instar larvae were dissected, trypsinized to single cells, and analyzed by FACS for their cell size by using the forward scatter (FCS). Expression of CycD/Cdk4 in a wild-type background led to an increase in the forward scatter of 20%-30%, which was reduced to 10%-15% in an hph02255/+ or Df(3R)3-4/+ background. Acridine orange staining showed that the suppression phenotype was not due to an increase in cell death. Furthermore, the cell cycle distribution was analyzed of eye imaginal discs from wandering larvae or pupae 48 hr after prepupae formation. At both time points, the increase in cells entering S and G2/M phases of the cell cycle due to ectopic expression of CycD/Cdk4 was not altered in an hph02255/+ background. Taken together, these data demonstrate that the cell size and proliferation functions of CycD/Cdk4 can be separated. Furthermore, Hph is required for the increase in cell size but not required for proliferation, suggesting that Hph functions downstream of CycD/Cdk4 in a growth-specific manner (Frei, 2004).
In the experiments described above, CycD/Cdk4 was induced in mostly postmitotic cells of the eye imaginal disc. To test suppression by hph in mitotically dividing cells, CycD/Cdk4 was induced during larval development, and wing discs cells were analyzed. Ectopic expression of CycD/Cdk4 shows a distinctive induction of growth: cells divide at a faster rate but are otherwise indistinguishable from control cells from the same disc. Therefore, when single clones are measured, the clone area is increased, and the clone consists of more cells with no change in cell size or cell cycle phasing. Since columnar cells of wing discs form a single cell layer, measuring the clone area gives an accurate estimation of the amount of mass that was accumulated during the growth of the clone. CycD/Cdk4 was overexpressed together with GFP in random clones using the hs-Flp Act>CD2>Gal4 system and analyzed after a 48 hr growth period. Compared to external control clones expressing only GFP, expression of CycD/Cdk4 caused a 75% increase in the median clone size. This phenotype depends on Hph, since the median clone size was reduced to control level in a heterozygous hph02255 mutant background. The suppression did not correlate with an increase in apoptosis, since coexpression of the cell death inhibitor p35 gave identical phenotypes. When cell size and cell cycle phasing were analyzed by FACS, there was no difference between cells expressing CycD/Cdk4 and internal control cells in either wild-type or hph/+ mutant backgrounds. These results demonstrate that the induction of growth by CycD/Cdk4 depends on normal levels of Hph. Furthermore, since hph suppressed growth but not proliferation in the eye imaginal discs, expression of CycD/Cdk4 in the wing should lead to a change in cell size if only growth but not proliferation were suppressed. However, no difference in cell size was detected, suggesting that the increase in proliferation caused by CycD/Cdk4 is secondary to the induction of growth (Frei, 2004).
Whether hph function was required for normal rates of cell growth was tested. Most hph0225/hph02255 or hph02255/Df(3R)3-4 animals die during embryogenesis, and only a few larvae hatch. These mutant larvae have severe growth defects and die within 2 to 3 days. Transheterozygotic hphS030304/hph02255 mutants develop normally until pupariation, but very few escaper adults eclose. These escapers are smaller than their heterozygous siblings but have normal body proportions. Weight measurements showed that hphS030304/hph02255 mutant flies are 18% lighter than heterozygotes. Therefore, hph mutant animals show a phenotype similar to homozygous cdk43 flies (Meyer, 2000) or wild-type flies reared at low oxygen (Palos, 1979; Frazier, 2001; Frei, 2004).
To test whether hph mutant cells are autonomously defective for growth, homozygous mutant clones were induced in the fat body using ionizing radiation. hph02255 was crossed to flies expressing GFP under the control of a constitutively active promoter inserted on the same arm of the chromosome as Hph (3R). The progeny were irradiated during embryogenesis, emerging larvae were grown in regular food, wandering third instar larvae were dissected, and their fat bodies were fixed and mounted. Homozygous hph02255 mutant cells lacked GFP, whereas heterozygous mutant cells expressed GFP. hph02255/hph02255 cells were smaller and contain less DNA than heterozygous neighboring cells. Importantly, the presence of a UAS-Hph transgene partially suppresses this phenotype, indicating that loss of Hph is the cause of the growth defect (Frei, 2004).
Whether ectopic expression of Hph is sufficient to stimulate growth was tested. EP3200 or UAS-Hph transgenes were used to induce Hph expression. Cell clones expressing Hph were induced in wing imaginal discs, and the median clone size was measured. Expression of Hph led to an increase in clone area very similar to CycD/Cdk4. Surprisingly, expression of Hph together with CycD/Cdk4 stimulated clonal growth to the same extent as Hph alone. However, in the presence of the apoptosis inhibitor p35, an additive phenotype was detected when both growth drivers were coexpressed. As for CycD/Cdk4, overexpressed Hph did not change cell size or cell cycle phasing, as assayed by FACS (Frei, 2004).
To test whether Hph functions downstream of CycD/Cdk4 also in this tissue, clones expressing Hph were induced in a homozygous cdk43 mutant background, and the median clone size was measured. Under these conditions, Hph led to a very similar induction of growth as in a wild-type background. FACS analysis indicates that there are no detectable changes in cell size or cell cycle phasing. These data show that Hph is sufficient to stimulate growth, and the finding that this stimulation is independent of Cdk4 suggests that Hph functions downstream of CycD/Cdk4 (Frei, 2004).
Ectopic expression of CycD/Cdk4 in the posterior compartment of the wing imaginal disc using the en-Gal4 promoter leads to an enlargement of the posterior compartment in adult wings with no change in trichome (hair) density. Since the trichome density is proportional to the number of cells per area, the increase in compartment size is due to more cells of the same size. When Hph was expressed under en-Gal4 control using EP3200 or UAS-Hph transgenes, a similar result was obtained: posterior compartments were bigger and contained more cells of the same size. Thus, Hph induces growth in a similar manner to CycD/Cdk4 in wing imaginal discs. However, when Hph was expressed in the eye imaginal disc using the GMR-Gal4 driver, no increase in cell size, as assayed by FACS, was observed. Furthermore, adult eyes were not enlarged. Further experiments are required to understand why Hph expression is not sufficient to increase growth in the eye imaginal disc (Frei, 2004).
Little is known about how Hph RNA or protein levels are regulated. In Drosophila embryos, hph is expressed uniformly and does not seem to be subject to patterning. Vertebrate cells have three Hph orthologs (Taylor, 2001): HPH-3 protein localizes to the nucleus, HPH-2 exclusively to the cytoplasm, and HPH-1 mainly in the cytoplasm with a little staining in the nucleus (Huang, 2002; Metzen, 2003; Frei, 2004).
To test whether the subcellular localization or levels of Hph are altered in response to CycD/Cdk4, polyclonal antibodies were raised to full-length Hph. To test the specificity of the antiserum, Hph was overexpressed in the posterior compartment of the wing using the en-Gal4 driver. Hph staining increased in posterior regions, both in the peripodial and columnar epithelium. Homozygous hph mutant cells, marked by the absence of GFP, lacked detectable Hph staining. Therefore, the serum is specific for Hph and is able to detect endogenous levels of Hph. Furthermore, in third instar imaginal wing discs, Hph staining was uniform throughout the disc and specific for the nucleoplasm of the cells. Very little staining was detectable in the cytoplasm or the nucleolus (Frei, 2004).
When Hph staining was analyzed in cells expressing ectopic CycD/Cdk4, increased Hph levels were observed. Antiserum staining in the peripodial epithelium shows that these cells had an increase in nucleoplasmic as well as cytoplasmic Hph. In contrast, when homozygous cdk43 mutant cells, marked by the absence of GFP, were analyzed, only background levels were detected. Therefore, Hph protein levels are regulated in response to CycD/Cdk4 gain and loss of function. When hph expression was analyzed by RT-PCR from wing discs expressing ectopic CycD/Cdk4, no effect on Hph RNA levels was detected. Furthermore, when whole third instar larvae were analyzed by RT-PCR or microarray analysis, no change in Hph expression was observed. Taken together, these observations suggest that CycD/Cdk4 affects Hph levels posttranscriptionally (Frei, 2004).
The hydroxylation activity of HPHs depends of Fe2+ bound to the active site. Therefore, iron chelators like deferoxamine mesylate (DFO) are commonly used to experimentally mimic hypoxic conditions. When Drosophila larvae are raised on regular food supplemented with 2 mM DFO, they show an induction in Hif-1α/β, as assayed with a reporter construct, that is very similar to that seen under hypoxic conditions. When CycD/Cdl4 or Hph are expressed in the posterior compartment of the wing imaginal disc, the increase in compartment areas is suppressed by DFO. Moreover, when DFO is added to flies expressing CycD/Cdk4 in the postmitotic eye using the GMR-Gal4 driver, the enlargement of the adult eye as well as the rough appearance is suppressed, however, not to the same extent as in the heterozygous hph mutant backgrounds. This suggests that the hydroxylation activity of Hph is required for its growth function and that Hph is a major growth effector of CycD/Cdk4 (Frei, 2004).
The finding that Drosophila Hph functions downstream of CycD/Cdk4 and is sufficient to increase growth when overexpressed suggests that CycD/Cdk4 and Hph work in a common pathway. Consistent with this, heterozygous hph mutants do not suppress the extra growth induced by components of the insulin pathway or dMyc. Moreover, increases in cell size and changes in cell cycle phasing induced by the insulin signaling pathway, dMyc, or Ras in wing imaginal disc cells, do not depend on Cdk4. Taken together, these results suggest that the CycD/Cdk4-Hph pathway functions separately from these other growth regulatory pathways (Frei, 2004).
Since the kinase activity of Cdk4 is required for the induction of growth and proliferation, Hph could be a phosphorylation target of Cdk4. The consensus sequence of vertebrate pocket proteins, the only known targets of cyclin D1/Cdk4, can be different from the classical CDK sequence. In Drosophila Rbf1, two potential sites have been found that disrupt its regulation by CycD/Cdk4 and CycE/Cdk2: T356PLTR and S728PHPK. Both sites are different from the vertebrate consensus sequence. Therefore, a search for putative consensus sequences on Hph is difficult. However, there are three sites that have the minimal requirement of a serine or threonine residue followed by a proline: T91PDAP, T204PGTT, and T285PPAA. None of these resemble the consensus sequences recognized by either the vertebrate or Drosophila complex. Nevertheless, future experiments should address whether CycD/Cdk4 phosphorylates Hph on these or other sites and how this affects Hph function (Frei, 2004).
It is proposed that in wing discs, Hph protein levels are regulated in response to CycD/Cdk4. Although it cannot be excluded that growth is also induced in an Hph-independent manner, the findings that overgrowth driven by CycD/Cdk4 and Hph is suppressed nearly completely by the iron chelator DFO or by heterozygosity for hph suggest that this is a major mechanism. Moreover, it is proposed that the small size of flies reared at low oxygen concentrations is caused at least partially by a decrease in Hph activity due to the absence of oxygen (Frei, 2004).
How does Hph induce growth? Since the hydroxylation inhibitor DFO suppressed the increases in growth caused by CycD/Cdk4 or Hph, Hph's hydroxylation activity is probably required. The only characterized hydroxylation target of Hph is Hif-1α, a mediator of the transcriptional response to hypoxia. Although mutant alleles of the Drosophila Hif-1α ortholog sima are not available, a partial loss-of-function allele of the Hif-1β ortholog, tango1, was available. To test the potential role of Hif1 in growth control, the ey-Flp/FRT method was used to generate flies in which the eyes were >80% homozygous mutant for tango1. If Hph stimulates growth by hydroxylating Hif-1α and targeting it for degradation, then loss of Hif-1 activity might be expected to result in overgrowth phenotypes. Contrary to this expectation, overgrowth was not observed in tango1/tango1 eyes. Moreover, GMR-driven expression of CycD/Cdk4 led to the same degree of overgrowth in tango1/tango1 eyes as in wild-type controls. Although these observations weigh against an important role for Hif-1 in Hph-driven growth, it is important to note that tango1 is not a null allele and that Tango is thought to be expressed in excess over its binding partner, Sima. Thus, further analysis using sima mutants and overexpression will be required to definitively test whether Hph drives cell growth via a Hif-1-dependent mechanism or through hydroxylation of novel targets. The finding that only one of the three vertebrate Hph orthologs is required for regulation of Hif-1α levels in vivo (Berra, 2003) further suggests that additional targets may be important (Frei, 2004).
There is little data that suggest a growth function for vertebrate HPH. Rat HPH-1/SM-20 was identified first as a gene upregulated by growth factors or serum. The induction is very fast and peaks at 60 min after stimulation. Remarkably, this induction does not require de novo protein synthesis, as it is not blocked by the translation inhibitor cyclohexamide. The effect on growth upon deregulation of mouse Falkor/HPH-3 is controversial: whereas expression of a C-terminal fragment induced cells to grow faster and to a higher density, expression of a wild-type construct had no effect. An antisense oligonucleotide specific for Falkor induced cells to grow faster. Thus, the function of vertebrate HPH family member in growth control is still ambiguous (Frei, 2004).
Drosophila Hph has at least two functions: response to hypoxia and regulation of growth. How are they linked? In response to hypoxia, Sima/Tango activity is strongly induced in endoreplicative tissues like trachea, gut and fat body, and to a much lesser extent, in imaginal discs (Lavista-Llanos, 2002). Although endopreplicative cells lacking Hph are impaired for growth, ectopic overexpression of Hph in these cells does not increase their size. In contrast, in imaginal discs, Hph can increase growth when overexpressed. It is speculated that in endoreplicative tissues, Hph's main function is to regulate the hypoxic response and, to a minor extent, growth, whereas in imaginal tissues, Hph's main function is to regulate growth. Taken to the environment of wild Drosophila, this suggests that hypoxic conditions, which are often found in fermenting fruit, may induce a strong hypoxic response in endoreplicative tissues. Since these tissues are metabolically highly active, this response may be required for the generation of sufficient ATP by the induction of glycolysis. In imaginal discs, cell cycle progression is not controlled primarily by extrinsic factors but by disc intrinsic growth cues. Therefore, even under hypoxic stress, growth and development of imaginal discs continues but may be slowed down, presumably by inactivation of Hph activity, in order to ensure the formation of adult animals (Frei, 2004).
In fat body cells, Hph is a nuclear protein, and homozygous Cdk4 mutant cells lack detectable Hph levels. Moreover, ectopic expression of CycD/Cdk4 leads to more Hph protein in the cytoplasm and/or the nucleus. Surprisingly, a reporter line showed an increase, rather than a decrease, in Sima activity upon expression of CycD/Cdk4. It is proposed that in the fat body, Hph induced by CycD/Cdk4 is not sufficient to hydroxylate Hif-1α. In addition to the cofactors oxygen and iron, hydroxylation activity requires the binding of 2-oxoglutarate to the active site of HPH (Epstein, 2001; Bruick, 2001). 2-oxoglutarate is an intermediate of the citrate cycle, and its levels might correlate with the metabolic activity of the cell. Therefore, Hph protein may be induced by CycD/Cdk4 but may require 2-oxoglutarate and oxygen for catalytic activity in the fat body. In this model, Hph would integrate the regulation of growth by CycD/Cdk4 and its upstream regulators, with the regulation of growth by the metabolic activity, mediated by oxygen and 2-oxoglutarate (Frei, 2004).
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.