Hypoxia induces a stereotypic response in Drosophila melanogaster embryos: depending on the time of hypoxia, embryos arrest cell cycle activity either at metaphase or just before S phase. To understand the mechanisms underlying hypoxia-induced arrest, two kinds of experiments were conducted. First, embryos carrying a kinesin-green fluorescent protein construct, which permits in vivo confocal microscopic visualization of the cell cycle, showed a dose-response relation between O2 level and cell cycle length. For example, mild hypoxia (Po2 ∼55 Torr) had no apparent effect on cell cycle length, whereas severe hypoxia (Po2 ∼25–35 Torr) or anoxia (Po2 = 0 Torr) arrested the cell cycle. Second, we utilized Drosophila embryos carrying a heat shock promoter driving the string (cdc25) gene (HS-STG3), which permits synchronization of embryos before the start of mitosis. Under conditions of anoxia, we induced a stabilization or an increase in the expression of several G1/S (e.g., dE2F1, RBF2) and G2/M (e.g., cyclin A, cyclin B, dWee1) proteins. This study suggests that, in fruit fly embryos, 1) there is a dose-dependent relationship between cell cycle length and O2 levels in fruit fly embryos, and 2) stabilized cyclin A and E2F1 are likely to be the mediators of hypoxia-induced arrest at metaphase and pre-S phase.
- cyclin A
- oxygen deprivation
- stabilization of expression
the cell cycle comprises several stages or phases that include gap phase 1 (G1), DNA replication (S phase), gap phase 2 (G2), and mitosis (M phase) (49). Movement through the various stages of the cell cycle is accomplished by two essential mechanisms: the sequential accumulation and degradation of several cell cycle regulatory proteins, such as the cyclins, and the sequential activation and inactivation of their cognate partners, the cdks, and other protein complexes (35, 48, 84). Therefore, there are two mechanisms of control of the cell cycle: levels of protein expression and the phosphorylation state of these proteins. The degradative pathways that operate during the cell cycle include the Skp1, cullin (cdc53), and F-box protein cdc4 (SCF) family of ubiquitin ligases at the G1/S transition and the anaphase-promoting complex (APC) at the G2/M transition, with some overlap. These families of ubiquitin-conjugating ligases, in conjunction with the proteosome 26S or cyclosome, mediate the destruction of agents such as cyclins A and B to permit the exit from metaphase (35, 82) and the degradation of several G1/S regulatory proteins such as E2F1, cyclin D, and cyclin E (36, 61), respectively.
Furthermore, the cell cycle is modulated by a series of intrinsic and extrinsic factors. Intrinsic control mechanisms include the oscillatory activity of protein complexes, such as the cyclin/cdk, that determine the periodicity of the cell cycle and checkpoint apparatus that can halt cell cycle activity in the event of an endogenous stress, such as incomplete DNA replication. The intrinsic factors also include mutations of several cell cycle genes (35, 48, 84). Extrinsic stimuli that can influence cell cycle progression include mitogenic factors, such as growth factors, and a variety of environmental stresses (see Refs. 22 and 42).
Environmental stresses can have direct effects on cell cycle activity and therefore on cell division and growth of organs and tissues, especially early in life. Ultraviolet (UV) and ionizing (IR) radiation is known to cause DNA damage that leads to cell cycle delay, cell cycle arrest, dysplasia, and/or cell death (8, 52, 60, 63). Other factors such as heat, overcrowding, and desiccation also can lead to cell cycle arrest (17). Oxygen (O2) deprivation or hypoxia is another such stress that appears to have a direct effect on cell cycle activity and growth (18, 30). Hypoxia can lead to either hypoplasia, as seen in the growth retardation reported in response to in utero hypoxia (23), or even hyperplasia, as demonstrated in the mouse pulmonary vasculature (29), as well as in the Drosophila tracheal system (34). Additionally, it has been demonstrated that hypoxia induces multiple effects on early rapidly cycling embryos of different species and can affect the cell cycle in a variety of ways (7, 18, 24, 56, 79). It can cause the cycle to halt completely, both reversibly (its activity resumes upon reoxygenation) and irreversibly (which leads to cell death) (2, 24, 81).
In the Drosophila embryo, the first 13 cell cycles are rapid and under the control of the products loaded by the mother. These early cycles exhibit only S and M phases. However, the zygotic control starts at cell cycle 14, and, more importantly, overexpression of String before cycle 14 will synchronize the embryos to the premitotic stage of cycle 14 (16).
Recently, we (13) and others (7, 12, 39, 56, 81) have demonstrated that hypoxia can lead to cell cycle arrest at either metaphase or pre-S phase in early cycling (stages 1–13) embryos. The question arises as to which specific mechanisms are employed by the cell in response to hypoxia to induce cell cycle delay or arrest. To understand the mechanisms involved in the hypoxia-induced metaphase and pre-S phase cell cycle arrests observed in fruit fly embryos, we have asked two questions. 1) Is there a hypoxic threshold for the arrest? 2) What proteins have altered levels of expression that could explain the hypoxia-induced arrest?
MATERIALS AND METHODS
Several antibodies were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development and maintained by the University of Iowa, Department of Biological Sciences (Iowa City, IA). These include monoclonal antibodies to β-tubulin (E7), cyclin A (A12), and cyclin B (F2F4), which were developed by M. Klymkowsky, C. F. Lehner, and P. H. O'Farrell, respectively. Rabbit polyclonal antibodies to cyclin A (Rb270) and cyclin B (Rb271) and a mouse monoclonal antibody to β-tubulin (Bx69) were also obtained from David M. Glover (Cambridge University, Cambridge, UK). Additionally, a mouse monoclonal antibody to cyclin B was provided by J. Raff (Wellcome/CRC Institute, Cambridge, UK), and a mouse monoclonal antibody to β-tubulin was obtained from Accurate Biochemical and Scientific (Westbury, NY). Antibodies to E2F1 were provided by T. Orr-Weaver (Whitehead Institute, MIT, Cambridge, MA) and by M. Asano of Duke University Medical Center (Durham, NC). The mouse monoclonal antibodies to E2F2 (Mei-8), RBF1 (DX-3, DX-5), RBF2 (DR6), and dDp (Yun-3, Yun-5) were kind gifts from N. Dyson (Massachusetts General Hospital Cancer Center, Harvard University, Boston, MA), and a rabbit polyclonal anti-dWee1 antibody was donated by S. Campbell (Univ. of Alberta, Edmonton, Canada). Rabbit polyclonal antibodies to cdk1 (cdc2) and String (cdc25) were generously provided by B. A. Edgar (Fred Hutchinson Cancer Research Center, Seattle, WA), and the mouse monoclonal antibody to Fizzy (cdc20) was a gift from I. Dawson (Molecular and Cellular Biology, Yale University, New Haven, CT).
Wild-type Canton-S strain containing the kinesin-green fluorescent protein (K-GFP) construct (Ma+-+-GFP2l2 [II]) was obtained from T. Xu (Yale Univ. School of Medicine, New Haven, CT). The heat shock String fly strain [w, Df(w)67; HS-STG3/TM6B] was provided by B. A. Edgar. The double-construct fly strain containing both K-GFP and HS-STG (w; K-GFP/K-GFP; HS-STG/TM6B) was generated in our laboratory.
Egg Collection and Preparation
Fly stocks were maintained at 25°C under standard conditions. Embryos were collected on apple juice plates supplemented with a small amount of yeast paste. Generally, adult fruit flies were anesthetized with 100% CO2, and, after a recovery period of 1–2 h, the first collection was discarded to eliminate old eggs from the sample population (16). To synchronize the eggs, two rapid 10-min collections were made and discarded. Subsequent collections for experimental purposes were made at 30 min, 1 h, 2 h, or 4 h after egg deposition (AED). Embryos were rinsed from the agar plates into a Nytex mesh holder with a solution containing 0.7% NaCl and 0.02% Triton X-100 (TX-100; Sigma, St. Louis, MO). They were then thoroughly rinsed with distilled deionized H2O (DDW). Embryos were then dechorionated in 50% domestic bleach (sodium hypochlorite; Chlorox, Oakland, CA) in DDW for 2 min and again rinsed thoroughly with DDW before any experimental manipulation.
Observing Living Embryos During Graded Hypoxia
Dechorionated K-GFP fly embryos were mounted onto coverslips, and each coverslip was placed into a specially designed perfusion chamber. The embryos were then enclosed within the perfusion chamber with a circular Thermanox plastic coverslip (15-mm diameter; Nalge Nunc, Naperville, IL), immediately perfused with normoxic insect medium (Shields and Sang M3 Insect Medium, Sigma) and viewed with a Bio-Rad MRC-1024 confocal microscope system (Bio-Rad Laboratories, Hercules, CA). Images were captured every 20–40 s from K-GFP embryos for one or two cell cycles under control normoxic conditions before switching to a hypoxic insect medium. The hypoxic medium had been previously bubbled with either 100% N2 (severe hypoxia, Po2 ∼25–35 Torr), 5% O2 (mild hypoxia, Po2 ∼55–60 Torr), or 2% O2 (moderate hypoxia, Po2 ∼40–45 Torr) for 2 h before the experiment and was continuously bubbled with the same gas throughout the experiment. The design of our perfusion chamber for in vivo confocal studies of embryos permitted us to study a dose response under mild, moderate, and severe hypoxia, as defined above. However, because of the open-ended design of the chamber, we could not reach anoxia (0 Torr) in these studies. Perfusion with the hypoxic solution was maintained for 30 min to 2 h. After the hypoxic period, perfusion with normoxic insect medium was reinitiated, and, on the resumption of cell cycle activity, embryos were observed for an additional one or two cell cycles during reoxygenation.
Immunoblot and Immunocytochemical Analysis on Drosophila Embryos
To generate a homogeneous population of embryos, we utilized a fruit fly strain carrying a heat shock protein 70 promoter-String (yeast homolog, cdc25) construct (HS-STG3) that permits embryonic cells, resting in G2, to be heat pulsed into an early but well-synchronized mitosis and subsequent S phase. After several precollections, as previously described, HS-STG3 embryos were collected on 100-mm apple juice plates over a 30-min period. The HS-STG3 embryos were then aged for 130 min after egg deposition (115–145 min AED) to generate embryos that had completed the first 13 cell cycle (nuclear) divisions and had entered the first G2 gap phase of Drosophila embryogenesis (cycle 14, Fig. 1). During the aging process, the embryos were dechorionated, returned to the agar plates, and subsequently placed in an incubator at 37°C for 40 min. A heat shock pulse of this duration induces global string mRNA and String (STG) protein expression in the embryo and synchronously drives all G2 cells into mitosis at cycle 14 within ∼5–10 min of returning the embryos to room temperature (RT). Mitosis is generally completed in 10–20 min, and an S phase that lasts for 45 min immediately follows this “premature” mitosis. The cells then return to G2 (cycle 14) and remain there for 70–170 min before resuming the normal developmental program (16). This was confirmed by in vivo observation of the K-GFP/HS-STG3 double construct (data not shown).
Synchronized, dechorionated HS-STG3 embryos were placed into a specially designed anoxia chamber that allowed the manipulation of the embryos while totally preserving the anoxic environment. The chamber was continuously gassed with 100% N2 through a single port. Two small escape ports were located on the other side of the chamber to allow for normalization of internal chamber pressure. O2 concentration was monitored continuously with a Clark-like O2 electrode (DO-051; Cameron Instrument, Aransas, TX). After calibration of the Clark electrode with alternating 100% N2 and room air (21% O2) gases, the chamber was gassed with 100% N2 (0% O2) for ∼30 min before insertion of the embryos submerged in Shield's and Sang medium in 15-ml centrifuge tubes into the chamber. Tubing connected to the N2 source was also placed directly into the 15-ml centrifuge tubes containing the embryos to ensure total anoxia in the microenvironment of the embryos. The Po2 in the anoxia chamber was maintained at 0 Torr, as continuously measured with the Clark O2 electrode, for the duration of the anoxic episode. After the heat pulse, embryos were held at RT for 4.5 or 12.5 min before anoxic exposure to recapitulate the arrests at metaphase (M) and pre-S phase (S), respectively (Fig. 1). Heat-pulsed embryos were then subjected to anoxia for either 5 min (MA1 and IB1) or 30 min (MA2 and IB2) to ascertain the effect of acute and protracted anoxia on protein expression (Fig. 1). At the end of the anoxic exposure, embryos were quickly removed and frozen in liquid N2 for immunoblot analyses. The appropriate controls were also run concurrently in medium maintained in room air or bubbled with 21% O2 (C1–C6). These embryos were then quickly frozen in liquid N2 for immunoblotting (n = 3). Alternatively, 100 embryos were hand-counted before being placed into the anoxia chamber on agar plates and subjected to the same protocol (n = 3). Data were pooled from these six experiments (n = 6).
For Western blotting, the embryos were homogenized with a glass-Teflon homogenizer (Thomas Scientific, Swedesboro, NJ) in Laemmli sample buffer (2% SDS, 10% glycerol, 67 mM Tris, 20 mM EDTA, 20 mM EGTA, 0.1% β-mercaptoethanol, pH 6.7) containing a cocktail of protease inhibitors. Protease inhibitors included a Complete tablet containing pronase, thermolysin, chymotrypsin, trypsin, and papain (Boehringer-Mannheim, Mannheim, Germany), a RIPA cocktail composed of pepstatin A (Sigma), leupeptin (Roche, Mannheim, Germany), and aprotinin (Boehringer-Mannheim), as well as phenylmethylsulfonylfluoride (PMSF; Sigma). Embryo samples were then assayed for protein content (BCA Assay; Pierce Chemical, Rockford, IL), and 10–20 μg of total protein in a 4× sample loading buffer (2% SDS, 10% glycerol, 67 mM Tris, 20 μM EDTA, 20 μM EGTA, 0.1% β-mercaptoethanol, pH 6.7, with a Complete tablet and PMSF) were loaded per lane for SDS-PAGE and subsequent transfer to polyvinylidene difluoride (PVDF) membranes (Immobilin-P, Millipore, Bedford, MA). Proteins were detected with secondary antibodies conjugated to horseradish peroxidase (Jackson ImmunoResearch Laboratories, Westgrove, PA, and Zymed Laboratories, San Francisco, CA) and the enhanced chemiluminescence system (Amersham, Little Chalfont, UK). Scanning densitometry of immunoblot films was performed on a Personal Densitometer SI scanner (Molecular Dynamics, Sunnyvale, CA) and analyzed using ImageQuaNT image analysis software (Molecular Dynamics).
Immunocytochemistry of Synchronized Embryos
To determine whether embryos utilized for immunoblot studies were arrested at the desired stages, i.e., metaphase and pre-S, embryos were subjected to the synchronization protocol and fixed for immunocytochemical processing. The detailed immunocytochemical procedure has been previously described (13). In brief, embryos were aged for 130 min AED, heat pulsed for 40 min at 37°C, and held at RT for 4.5 min (metaphase arrest) and 12.5 min (interphase/pre-S arrest) before being exposed to anoxia for 30 min. After the anoxic exposure, embryos were fixed inside the anoxia chamber with 2% paraformaldehyde in PBS and subsequently processed for the immunocytochemical detection of α-tubulin and propidium iodide.
Data are reported as means ± SE, and results were analyzed using the Wilcoxon signed-rank test and one-way ANOVA. Differences in the means were considered statistically significant when P < 0.05.
Effect of Graded Hypoxia on Drosophila Melanogaster Embryonic Cell Cycle Kinetics
The exposure of the embryos to mild hypoxia (5% O2) appears to have a limited impact on cell cycle duration and kinetics during the first 13 cycles (see Fig. 2B, a–f) and looked very similar to control embryos under normoxic conditions (Fig. 2A, a–f). Furthermore, energids (nuclei and associated cytoskeletal elements within the embryonic syncytium) within embryos did not show any evidence of abnormalities or deranged movement of intracellular organelles or other structures. Formation of the spindle apparatus and centrosomes was not impeded nor was cytokinesis or the formation of asters. Subsequent cell cycles proceeded normally. Centrioles replicated and migrated toward the poles of nuclei for the next round of DNA replication and mitosis. Switching to a normoxic perfusate after 5% O2 did not have an effect on cell cycle activity. Moderate hypoxia (2% O2), by contrast, had a dramatic effect in terms of kinetics and duration (Fig. 2B, g–l). Within a few seconds of switching to a perfusate bubbled with 2% O2, energids simultaneously and synchronously slowed their rate of cell cycle progression and thereby markedly protracted cell cycle length. Cell cycle duration, which is normally ∼950 s, increased by approximately fivefold and extended to ∼4,300 s (70 min), with no bias to any particular cell cycle stage (Fig. 2C). The morphology of the subcellular organelles and structures, i.e., spindle apparatus and centrosomes, appeared to be unaffected, as during 5% O2 or normoxic exposure. A complete cessation of cell cycle activity did not occur. Therefore, embryos slowed their cell cycle activity tremendously, with no other abnormalities observed. Switching back to a normoxic perfusate allowed the embryos to rapidly resume their original cell cycle program. A complete cell cycle arrest occurred only when very low O2 concentrations (severe hypoxia, <0.5%) were used in these embryos (Fig. 2A, g–l). Depending on the time of initiation of severe hypoxia, embryos either arrested at metaphase or arrested just before the S phase. Embryos that were exposed to severe hypoxia just before metaphase halted their cell cycle activity at metaphase with chromosomes aligned along the metaphase plate and a fully developed spindle apparatus (data not shown). On the other hand, embryos that were exposed just after metaphase or during or past anaphase, migrated through the rest of the cycle and arrested in a hypoxia-induced pre-S phase (Fig. 2A, g–l). Furthermore, the normoxia and hypoxia treatments gave the same result as before (13).
Cell Cycle Regulatory Protein Expression During Anoxia in Drosophila Embryos
Synchronization of Drosophila embryos for immunoblot studies.
Embryos were aged, heat-pulsed, and maintained at RT for 4.5 and 12.5 min before anoxic exposure to recapitulate the metaphase and interphase/pre-S arrests, respectively. They were then fixed and processed for the immunocytochemical localization of α-tubulin and propidium iodide to detect the plasma membrane and chromatin material of cells/energids in the embryos, respectively, to determine the cell cycle stage at which the embryos arrested. Embryos held at RT for 4.5 min after the heat pulse (AHP) before the anoxic exposure were seen to be arrested at metaphase (data not shown), whereas embryos held for 12.5 min at RT AHP before anoxia were arrested at interphase (see Fig. 3B). All of the cells/energids in these embryos were at the same cell cycle stage with condensed, centrally located chromosomes predictive of interphase. Control embryos (Fig. 3A) were also in interphase as predicted, since they would have completed a full cell cycle and returned to the interrupted G2 after 42.5 min AHP.
Pre-S phase arrest and G1/S proteins.
To look for the mechanisms underlying the pre-S phase cell cycle arrest, we examined several G1/S proteins by Western blotting to determine their levels in response to low O2 (Fig. 4A). dE2F1, which is the major positive regulator of the initiation of S phase and DNA replication, was detected as a single band of 90–95 kDa. dE2F1 significantly increased over control levels by ∼45% after 5 min of anoxia (Fig. 4B, B1 vs. C5) in embryos that were arrested during pre-S phase. At the end of the anoxic period (30 min), dE2F1 levels did not change or declined somewhat but were still greater over its equivalent control (Fig. 4B, B2 vs. C6). A similar response pattern was seen for dE2F2 (data not shown) during the pre-S phase arrest.
RBF proteins (RBF1, RBF2) are considered negative regulators of the G1/S transition in Drosophila (14). The responses of RBF1 (data not shown) and RBF2 (Fig. 4C) were similar to the dE2F1 anoxic response (Fig. 4B). RBF2 appears as a band of ∼83 kDa on Western blot. The expression of RBF2 was significantly increased (60%) over comparable controls (Fig. 4C, B1 vs. C5) and then declined over the anoxic period but not below control levels (Fig. 4C, B2 vs. C6). In summary, G1/S regulatory proteins examined in this study demonstrated, in general, an increase in expression during the early stages of anoxia but subsequently declined over the next one-half hour of anoxic exposure.
Pre-S phase arrest and G2/M proteins.
Cyclins A and B are necessary for the G2/M transition, as well as progress through the initial stages of mitosis. However, recent studies using microarray analyses have indicated that cell cycle regulatory proteins that have traditionally been ascribed specific roles in cell cycle progression may very well regulate other stages in the cell cycle (see Ref. 47). We therefore examined the response of proteins known to control the G2/M transition to determine their role at the G1/S transition, i.e., in the pre-S phase arrest paradigm. Cyclin A is detected as a band of ∼59 kDa, and cyclin B and dWee1 kinase are detected as single bands of ∼69 kDa and ∼110 kDa, respectively (Fig. 5). During the pre-S phase arrest, cyclin A, cyclin B, and the dWee1 kinase either stayed close to baseline or increased significantly over control (Fig. 5, B1 vs. C5). Similar to the expression profile of dE2F1 (Fig. 4B) and RBF2 (Fig. 4C) during the pre-S phase arrest, the expression of cyclin A and dWee1 kinase declined but remained above comparable controls (Fig. 5, B and D, B2 vs. C6) during the anoxic exposure. On the other hand, protein expression levels of cyclin B continued to increase during anoxia such that cyclin B expression in embryos was significantly greater (40%) than that seen in control embryos (Fig. 5C, B2 vs. C6) at the end of the 30-min anoxic exposure.
Metaphase arrest and G2/M proteins.
Regulation of the G2/M transition and progress through mitosis are controlled by a relatively complex hierarchy of cell cycle proteins, including the major positive regulators cyclins A and B and the histone H1 kinase (cdc2 or cdk1 in yeast), as well as the negative and positive regulators of cdk1 activity, such as dWee1 and String (cdc25 in yeast), respectively. During an anoxia-induced metaphase arrest, the expression of several cell cycle regulatory proteins such as cdk1, dDp, dE2F2, Fizzy (cdc20 in yeast), RBF1, and String was either unaltered or slightly increased compared with relevant controls (data not shown). Interestingly, the cyclins A and B (Fig. 6) also demonstrated stabilization in expression (Fig. 6, B and C, C2 vs. A1) during a metaphase arrest. Even though these changes were not significant, it is known that cyclins A and B must be degraded during the metaphase-to-anaphase transition of mitosis. Therefore, it is apparent that cyclins A and B are not degraded during anoxia, since control and anoxic embryos (Fig. 6, B and C, C2 vs. A1 and A2 vs. C3) have equivalent expression levels throughout the anoxic exposure.
Metaphase arrest and G1/S proteins.
The examination of G1/S proteins during the metaphase arrest also revealed that, in general, cell cycle protein expression was stabilized during anoxia. For example, the expression of dE2F1 during anoxia was equivalent to control throughout anoxia (data not shown). On the other hand, the expression of RBF2 appeared to be lower in anoxia-exposed embryos compared with controls at both anoxia time points examined (data not shown).
Our data indicate that changes in the level of O2 to which D. melanogaster embryos are exposed have a major impact on cell cycle progression. Moderate hypoxia (40–45 Torr, 2% O2) slowed cell cycle progression almost fivefold compared with control or mild hypoxia (55–60 Torr, 5% O2), whereas severe hypoxia (25–35 Torr, 100% N2) resulted in total cell cycle arrest at two points: metaphase and pre-S phase. In our current studies, we have also investigated whether, during these arrests, analysis of protein levels can be useful for understanding the basis of the hypoxia-induced phenomena.
Effect of Low O2 on Cell Cycle Progression
The mechanisms that are potentially involved in slowing the cell cycle in response to a hypoxic stress are currently not known, although a drop in ATP levels is one possibility (24). Depletion of ATP stores during hypoxia may slow the cell cycle of D. melanogaster embryos by decreasing overall metabolic activity. Indeed, it has been hypothesized that a reduction of the metabolic rate will decrease protein and RNA synthesis and metabolism, thus slowing down the cell cycle (7 and reviewed in Ref. 74). Decreased protein synthesis in hypoxia has been observed in a variety of species, including marine snails, fish, turtles, and humans (see 39). Although this may still be true in Drosophila embryos, the decreased protein levels are not homogenous. For example, we noted an increase in cyclin B levels during the anoxic period, indicating that active protein synthesis or a possible decrease in protein degradation occurs during anoxia in Drosophila embryos.
O'Farrell and colleagues (12, 81) have suggested another possibility: nitric oxide, rather than ATP, plays a critical role in hypoxia-induced arrest in Drosophila. Yet other investigators believe that phosphorylation/dephosphorylation, and, in particular, phosphorylation of histone in Caenorhabditis elegans (56) and eIF-2α in Littorina littorea (39), is a requisite component of the hypoxia checkpoint. Therefore, it is possible that several cellular mechanisms, acting in a concerted manner, are involved in the hypoxia-induced slowing and arrest of cell cycle activity.
Role of Cyclin A and B Proteins in Transducing an Anoxia-Induced Arrest
Because we did not observe a decrease in cyclin A and cyclin B levels during both metaphase and pre-S phase anoxia-induced arrests, the question arises whether the stabilized cyclin A or cyclin B can induce arrest. In general, G1 arrest can occur in response to DNA damage elicited by ionizing radiation and UV radiation (1), whereas G2 arrest can be induced by the spindle checkpoint and/or DNA damage (see Ref. 8). Additionally, mutations in cyclins A and B have been shown to induce temporary or permanent arrest (41, 48, 51).
In Drosophila, cyclin A is normally degraded before the metaphase-to-anaphase transition, cyclin B is degraded at the metaphase-to-anaphase transition, and cyclin B3 is degraded during anaphase (11, 41, 59 and reviewed in Ref. 75) by the mitotic ubiquitin ligase system, the APC (82). Because the degradation of cyclin A and other proteins is required for cells to progress beyond metaphase (62, 82), the lack of reduction of cyclin A levels (e.g., inhibition of cyclin A proteolysis) may induce an arrest at the metaphase-to-anaphase transition (9, 77). To date, there is significant evidence that stabilized cyclin A can induce a metaphase arrest. Rimmington et al. (66) and Sigrist et al. (69) were among the first to report that stable cyclins A, B, and B3 result in a delay at metaphase, early anaphase arrest, and a late anaphase arrest, respectively. More recently, it has been demonstrated that nondegradable cyclin A induces an arrest at metaphase in Drosophila and mammalian cells (21, 33, 59) and an arrest at metaphase with condensed chromosomes in Xenopus embryos (45). Stable cyclin B and cyclin B3 induce arrests during anaphase B (58) and in late anaphase with deep cytokinetic furrows (59), respectively.
Furthermore, Su and Jaklevic (76) have recently reported that DNA damage-induced stabilization of cyclin A is responsible for the DNA damage-induced cell cycle arrest at G2 in Drosophila gastrula cells and that cells lacking cyclin A do not delay in metaphase in response to DNA damage. In addition, overexpression of XDRP1, a Xenopus dsk2-related protein important in the spindle checkpoint pathway that leads to inhibition of the degradation of cyclin A (but not cyclin B), blocks embryonic cleavage cycles (20). Therefore, Parry and O'Farrell (59) suggest that cyclin A-like cyclins act to inhibit the metaphase-to-anaphase transition in diverse species and that degradation of these cyclins is required for progress through the cell cycle.
Although our results indicate that both cyclins A and B are stabilized during anoxia, the morphological characteristics of our embryos, which display a metaphase arrest profile with condensed chromosomes on the metaphase plate (12, 13), leads us to believe that cyclin A, rather than cyclin B, is the major player in the metaphase arrest, since stable cyclin B would induce an arrest at the end of anaphase, not at metaphase. Clearly, this does not preclude stable cyclin B from participating in the extended arrest. Because cyclin B is stabilized by the spindle checkpoint but cyclin A is not (11, 80), the present study indicates that anoxia may activate both the spindle checkpoint as well as a separate checkpoint subserved by cyclin A.
The difference between our previous study (13), in which we report a lower level of cyclin B with anoxia, and the present, in which we report a stable or increased cyclin B, is due primarily to 1) the cyclical nature of cyclin B during mitosis and 2) the time at which the embryos were fixed in relation to the cyclical nature of cyclin B. Cyclin B is higher in control embryos that have not passed mitosis than in anoxic embryos that have progressed beyond mitosis, whereas anoxic embryos that have passed mitosis and arrested in interphase have similar or higher levels of cyclin B than control embryos at a similar or more advanced stage.
It is conceivable also that cyclin A could induce arrest of cell cycle activity by other mechanisms. Cyclin A has been reported to inactivate cdh1, or Fizzy-related, which is responsible for the activation of the APC via phosphorylation and dissociation of cdh1 from the APC (55, 70) and which permits egress from metaphase. The other intriguing possibility is that cyclin A can also participate in a G1/S arrest via its role in S phase progression by suppression of E2F1 activity (37). For example, coexpression of cyclin A/cdk1, with an absolute requirement for cyclin A and the simultaneous accumulation of E2F, has been demonstrated to induce an S phase arrest, possibly by blocking E2F in Drosophila larval cells (26, 27, 46). We propose that stabilization of cyclin A is necessary and sufficient to induce a metaphase arrest during anoxia.
Role of dE2F1 and RBF2 in Anoxia-Induced Arrests
In mammalian cells, the transcription factor E2F1 is important for the initiation of the S phase in moving from the quiescent (G0, G1) to the proliferative state (reviewed in Ref. 15). E2F regulates genes involved in the initiation of DNA replication (cdc6), cell cycle activity (cdc25A, cyclins A and E), and growth (bMyb, p107, E2F1). Although there are six E2F family members (E2F1–6) in mammals, only two members have been described in Drosophila, dE2F1 and dE2F2 (reviewed in Ref. 15; 47, 68, 78).
E2F1 accumulates in late G1 phase of the cell cycle. It is then rapidly degraded in S/G2 phases of the cell cycle via interaction with the F-box protein p45SKP2, a component of the ubiquitin ligase SCFSKP2 (reviewed in Ref. 36; 46, 61). dE2F1, like E2F1, -2, and -3 in mammals, activates, and dE2F2, akin to mammalian E2F4, -5, and perhaps -6, represses gene transcription, while dE2F1 and dE2F2 antagonize each other (4, 6, 19, 68, 78). Surprisingly, in the present study, anoxia induced an increase in the expression of dE2F1 during the pre-S phase block. Although the anoxia-induced increase in dE2F1 protein expression was unexpected, it is not unparalleled: O'Connor and Lu (54) have reported that hypoxia of 24-h duration induced an increase in the expression of E2F1 that was independent from both p53 and pRb. Additionally, other stresses such as UV, IR, and DNA-damaging agents such as cisplatin and etoposide are also able to increase the levels of E2F1 in a variety of tumor cells (27, 31, 44, 54). Although it is known that the overexpression of E2F1 in normal cells leads to S phase entry and apoptosis (reviewed in Ref. 10), this was not observed in our study. However, it has been noted that E2F1 can also function as a repressor during the cell cycle (4, 6, 19).
Several studies implicate E2F1, and possibly -2 and -3, in repressing gene activity to permit exit from the cell cycle (6, 67). For example, Zhang et al. (83) believe that active repression by E2F/Rb and not inactivation of E2F is responsible for the G1 arrest triggered by contact inhibition, p16INK4a, and transforming growth factor (TGF)-β. Additionally, He et al. (27) propose that E2F actually suppresses the onset of S phase in cells arrested by γ-irradiation. A direct role for E2F1 in the DNA damage checkpoint-induced cell cycle arrest is also suggested by its higher levels after DNA damage via protein stabilization mediated by ATM-induced phosphorylation (44, 54). Furthermore, a stabilized form of E2F1 has been demonstrated to induce an S phase delay and apoptosis (38). However, although E2F has an essential role in cell cycle arrest in several cell lines (83), this stress-induced E2F1 does not appear to be transcriptionally active but yet plays a role in DNA damage-induced G1 arrest (54).
Recently, E2F1 has been shown, via microarray analysis, to possibly play a nontraditional role at G2/M (4, 32, 63, 65, 71). These studies implicate E2F in mitosis, chromosome segregation, the spindle checkpoint, DNA repair, chromatin assembly/condensation, apoptosis, differentiation, and development (32; reviewed in Ref. 47). For example, Polager and Ginsberg (63) have demonstrated that E2F1 is essential for maintaining a genotoxin (doxorubicin)-mediated G2 arrest by repressing critical mitotic regulators such as stathmin (oncoprotein 18) and aurora and Ipl-1-like midbody-associated protein kinase. Therefore, anoxia-induced E2F1 protein may participate, in addition to the pre-S phase arrest, in the arrest at metaphase.
The retinoblastoma protein (pRb) family members are critical regulators of cell cycle activity and play a predominant role in the regulation of E2F activity (3, 14, 42). Our study indicates that RBF1 is stabilized (data not shown) and that RBF2 is elevated during the anoxia-induced pre-S phase arrest. RBF1 is known to participate in a DNA damage-induced G1/S arrest (25, 43), as well as during TGF-β- and p16ARF-induced arrests (83). RBF2 is a newly described protein in Drosophila that has homology to human pRb and other pocket proteins such as RBF1, p130, and p107. In conjunction with dE2F2, RBF2 has been proposed to have global repressor function, binding to and inhibiting several promoter sites in dE2F1-responsive genes, antagonizing dE2F1/dDp transactivation, and, when overexpressed, blocking S phase entry in vivo (50, 72, 73). Stevaux et al. (72) indicated that RBF2 can function as a repressor, either alone or in combination with the E2F family (19). Therefore, we raise the question as to whether the increase in RBF2 noted in these studies could participate in an anoxia-induced cell cycle arrest at G1/S.
Role of Other Proteins Such as dWee1 Kinase in Anoxia-Induced Arrests
The dWee1 and dMyt1 kinases are known to be involved in regulation of the G2/M transition in Drosophila via their inhibitory phosphorylations on the mitotic activator cdc25 (String) (5, 28, 57). dWee1 and dMyt1 are active during S phase to prevent a premature mitosis, can be activated during the DNA damage checkpoint to arrest cells in G2 (40; reviewed in Ref. 53), and have been suggested to block apoptosis via interactions with p53 (64). However, in our study, an increase in the expression of dWee1 was seen during the pre-S phase arrest in HS-STG3 embryos. More experiments are required to elucidate the possible activities of dWee1 at points in the cell cycle other than at the G2/M transition.
In summary, several stresses are capable of invoking the checkpoint apparatus, whether it is DNA damage, failure of DNA replication, or spindle damage. However, it appears that hypoxia stimulates cell cycle arrest via redundant, known checkpoint pathways but also by unique and independent pathways. For example, whereas the spindle checkpoint induces degradation of cyclin A, hypoxia induces stabilization of cyclin A; therefore, multiple pathways are likely to be invoked during hypoxia (60). Analysis of the interactions of cell cycle proteins during hypoxia should begin to shed light on pathways that are exclusively activated by hypoxia compared with those activated by other stresses such as DNA damage.
T. Xu was supported by National Cancer Institute Grant CA-69408, the Pew Scholar Award, and Howard Hughes Medical Institute; G. G. Haddad was supported by National Institutes of Health Grants T32-HL-07778, P01-HD-32573, and NS-35918; and R. M. Douglas was supported by a Parker B. Francis Fellowship.
We thank Shelagh Campbell, Iian Dawson, Nick Dyson, Bruce A. Edgar, Haig Kesheshian, and Amanda Purdy for insightful discussions related to this project and Ralph Garcia and Joe Zavilowitz for superb technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society