The Drosophila embryo is a promising model for isolating gene products that coordinate S phase and mitosis. We have reported before that increasing maternal Cyclin B dosage to up to six copies (six cycB) increases Cdk1–Cyclin B (CycB) levels and activity in the embryo, delays nuclear migration at cycle 10, and produces abnormal nuclei at cycle 14. Here we show that the level of CycB in the embryo inversely correlates with the ability to lengthen interphase as the embryo transits from preblastoderm to blastoderm stages and defines the onset of a checkpoint that regulates mitosis when DNA replication is blocked with aphidicolin. A screen for modifiers of the six cycB phenotypes identified 10 new suppressor deficiencies. In addition, heterozygote dRPA2 (a DNA replication gene) mutants suppressed only the abnormal nuclear phenotype at cycle 14. Reduction of dRPA2 also restored interphase duration and checkpoint efficacy to control levels. We propose that lowered dRPA2 levels activate Grp/Chk1 to counteract excess Cdk1–CycB activity and restore interphase duration and the ability to block mitosis in response to aphidicolin. Our results suggest an antagonistic interaction between DNA replication checkpoint activation and Cdk1–CycB activity during the transition from preblastoderm to blastoderm cycles.

---

EMBRYONIC development, except in mammals, is initiated by cleavage stages that are distinguished by rapid and "synchronous" cell cycles that are controlled by maternal gene products. Characteristically, these early cycles have only S and M phases (MURRAY and HUNT 1993). Both biochemical and morphological studies have documented that sequential activation and inactivation of Cdk1–CycB play a major role in regulating these biphasic cycles (EDGAR et al. 1994, 1995; SU et al. 1998; JI et al. 2004).

In Drosophila embryos, the first 13 mitoses, which occur without cytokinesis and thus produce a syncytium, are maternally controlled. Cell cycle length is not different during the first 6 of these cycles (9 min/cycle at 21°). In subsequent cycles, interphase length gradually extends (i.e., mitosis becomes delayed) from 200 sec in cycle 6, to 300 sec in cycle 10, to 900 sec in cycle 13 (YU et al. 2000; JI et al. 2004). In addition, metaphase and anaphase, but not interphase durations, differ in different regions of the embryo, resulting in metasynchrony that correlates with local variation in Cyclin B (CycB) concentration (JI et al. 2004). Metasynchrony is also observed during the blastoderm cycles (cycles 10–13), when interphases become increasingly longer; however, at these stages the entire cell-cycle duration differs within the embryo. Nuclei at the poles divide faster (FOE and ALBERTS 1983; JI et al. 2004), which correlates with their lower nuclear densities (YASUDA 1992; BLANKENSHIP and WIESCHAUS 2001). Thus, metasynchronies at blastoderm are not a result of propagating mitotic waves.

We previously proposed that interphase extension in preblastoderm cycles (before cycle 10) occurs because CycB becomes limited. This conclusion is based on the following three observations: First, in later preblastoderm cycles (closer to cycle 10), when interphases become longer, embryos have less CycB than in cycle 6, presumably due to local degradation during each metaphase/anaphase transition (EDGAR et al. 1994; HUANG and RAFF 1999; STIFFLER et al. 1999). That is, CycB levels correlate inversely with interphase length. Second, increased CycB delays the onset of interphase extension; interphase extension occurs earlier in embryos from mothers with one copy of the cycB gene ("one cycB"; wild type has two copies) and later in embryos from mothers with four or six copies of cycB ("four cycB" or "six cycB" embryos; JI et al. 2002, 2004). Third, embryos from mothers lacking negative regulators of Cdk1–CycB, such as the DNA-replication checkpoint gene grapes (grp), which encodes a homolog of Chk1 kinase, exhibit interphase extensions not different from wild type during preblastoderm cycles (SIBON et al. 1997; JI et al. 2004). This is consistent with the idea that CycB limitation, rather than DNA replication checkpoint, is responsible for lengthening of interphase (i.e., delaying mitosis) before cycle 10 (JI et al. 2004).

Nuclei migrate outward during later preblastoderm cycles and reach the embryo periphery by the end of cycle 10 to form a blastoderm. In contrast to preblastoderm cycles, interphase lengthening in blastoderm cycles (cycles 10–13) appears to be regulated both by Cyclin B levels and by the DNA replication checkpoint (SIBON et al. 1997; JI et al. 2004).

Homozygous mutants in DNA checkpoint genes, mei41 (atr), grp (dChk1), or mus304 (dATRIP) are viable but embryos from such mothers extend interphase length to a lesser degree than in wild type during blastoderm cycles (SIBON et al. 1997, 1999, 2000; BRODSKY et al. 2000; YU et al. 2000; JI et al. 2004). The finding that Grp is needed for interphase lengthening in blastoderm but not preblastoderm cycles raises the possibility that DNA replication checkpoint is not functional before cycle 10. This has not been tested experimentally.

In addition, to delay interphase extension an increase of maternal cycB gene doses also leads to correspondingly more Cdk1–CycB activity, shorter microtubules, and slower nuclear movement (STIFFLER et al. 1999; JI et al. 2002). This is consistent with the fact that Cdk1–CycB activity regulates microtubule dynamics (STIFFLER et al. 1999). Thus, in embryos with elevated Cdk1–CycB levels, nuclear migration, a microtubule-dependent process, is impaired and not all nuclei are at the periphery by cycle 10 (JI et al. 2004). In addition, in interphase of cycle 14, we observe divisions that do not occur in wild type (two cycB) and uneven nuclear distributions. Remarkably, even with a total of six maternal cycB copies, offspring develop normally with adult hatching frequencies not different from those of controls. We used the nonlethal phenotypes at cycles 10 (nuclear migration) and 14 (abnormal mitosis) in six cycB embryos to perform a dominant-modifier genetic screen to identify maternal proteins that interact with Cdk1–CycB in regulating preblastoderm and blastoderm cycles.

This screen described three classes of suppressors: those that suppressed both cycle 10 and cycle 14 phenotypes, those that suppressed only at cycle 10, and those that suppressed only at cycle 14 (JI et al. 2002). Suppressors in the first two classes included regulators of microtubule dynamics and the metaphase–anaphase transition processes. The latter are of particular interest because it is in anaphase where the nuclear and cytoskeletal cycles are coordinated for proper separation of the chromatids (JI et al. 2005). Suppressors in the last class are candidates for the regulation of CycB function in the DNA replication checkpoint.

Here, we expanded our original screen for suppressors of six cycB phenotypes by using 104 molecularly and 11 cytologically defined deficiencies from the Exelixis (PARKS et al. 2004), DrosDel (RYDER et al. 2004), and Bloomington collections. This increased genome coverage from 70 to 83% and identified dRPA2 as a specific suppressor gene of the cycle 14 phenotype. In addition, we used aphidicolin (aph.), a chemical inhibitor of DNA synthesis, to demonstrate that a DNA replication checkpoint becomes detectable only during blastoderm cycles. Using embryos with different maternal CycB dosage, we demonstrate that increases in CycB delayed the time at which the DNA replication checkpoint could be detected. Furthermore, a reduction in dRPA2 also suppressed the later detectable DNA replication checkpoint in six cycB embryos. These results can be explained in a model in which lowered dRPA2 levels activate Grp/Chk1 to counteract excess Cdk1–CycB activity and restore interphase duration and the ability to block mitosis in response to aph.

Fly stocks:

A Sevelen stock was used for two cycB (+) control, as well as for the genetic background for all experimental stocks. Females carrying either four extra copies of cyclin B or one copy of cyclin B (JACOBS et al. 1998) produced embryos referred to as six cycB and one cycB, respectively. Deficiency (Df) stocks were obtained from the Bloomington stock center, from the Szeged stock center, or as previously described in JI et al. (2002). CG9273 (referred to as dRPA2¹) mutant stock, yw; P[y⁺ w⁺] CG9273^KG00759/CyO; +/+, was obtained from Bloomington. Nomenclature of genotypes is as follows: For deficiency screen, "Df/two cycB" embryos were from Df/w; cycB⁺/cycB⁺; +/+ or w/+; Df, cycB⁺/cycB⁺; +/+ or w/+; cycB⁺/cycB⁺; Df/+ females; "Df/six cycB" embryos were from Df/w; 2P[w⁺ cycB⁺], cycB⁺/cycB⁺; 2P[w⁺ cycB⁺]/+ or w/+; Df, cycB⁺/2P[w⁺ cycB⁺], cycB⁺; 2P[w⁺ cycB⁺]/+ or w/+; 2P[w⁺ cycB⁺], cycB⁺/cycB⁺; Df/2P[w⁺ cycB⁺] females. For live imaging, "one cycB" refers to embryos from w/+; cycB^–/His2AvD-EGFP, cycB⁺; +/+ females, "two cycB" to embryos from w/+; cycB⁺/His2AvD-EGFP, cycB⁺: +/+ females, "six cycB" to embryos from w/+; 2P[w⁺ cycB⁺], cycB⁺/ His2AvD-EGFP, cycB⁺; 2P[w⁺ cycB⁺]/+ females, "dRPA2/six cycB" to embryos from w/+; dRPA2^–, 2P[w⁺ cycB⁺], cycB⁺/ His2AvD-EGFP, dRPA2⁺, cycB⁺; 2P[w⁺ cycB⁺]/+ females, and "dRPA2/two cycB" to embryos from w/+; dRPA2^–, cycB⁺/ His2AvD-EGFP, dRPA2⁺, cycB⁺; +/+ females. His2AvD-GFP was used only for live analysis.

Embryo collection and immunohistochemistry:

For egg collections, flies were put on fresh food for 1 hr at 25° and then transferred to grape-juice-containing agar plates with yeast paste for two 30-min egg collections. For the genetic screen we collected and fixed embryos as described previously (JI et al. 2002). Embryos were fixed and stained at two stages, cycles 10 and 14. After precollections, which were used to remove over-aged embryos, collections of embryos for cycle 10 analyses were made for 30 min and incubated for 90 min from the midpoint of the collection before fixation. Collections of embryos for cycle 14 analyses were made for 60 min and incubated for 150 min from the midpoint of the collection before fixation. Embryos were dechorionated in 100% bleach for 1.5 min and washed for 2 min alternately with 0.1% PBS-Tx and distilled water. Antibodies to histone H1 (1:300; Chemicon, Temecula, CA) and phosphorylated histone H3 (1:300; Upstate Biotechnology, Lake Placid, NY) were used.

Generation of the genomic His2AvD-EGFP construct:

Since the enhanced-GFP variant is about seven times more intense than the GFP-S65T variant, we modified the original His2AvD-GFP transposon (CLARKSON and SAINT 1999). The genomic His2AvD-EGFP fusion gene was generated by modifying the pONIXAvDGFP-S65T plasmid (CLARKSON and SAINT 1999). The plasmid pONIXAvDGFP-S65T carries a 4.1-kb genomic His2AvD fragment with GFP-S65T at the end of the exon 4 in a pCaSpeR4-based vector (CLARKSON and SAINT 1999). To substitute GFP-S65T with EGFP, we first used PCR to amplify the following two fragments, flanked by two XhoI restriction endonuclease sites, from the pONIXAvDGFP-S65T plamid: One fragment contains exon 4 and part of intron 3 (5'-CCAAGCTTGTCGACGCGCACATTATCAGTTG-3' and 5'-GAAGATCTGCCTGCGACAGAATGACG-3' as primers); the other fragment contains the 3'-untranslated region (3'-UTR) and part of the pONIX vector (5'-GCCAGTCGGCAATCGGACGCC-3' and 5'-GCGTCGACTCTAGATACGACACACAGCGCAACATC-3' as primers). These two fragments were then separately subcloned into the pGEM-T Easy vectors (Promega, Madison, WI) and verified by sequencing. The EGFP fragment was cut out from the pEGFP-N1 plasmid (Clontech Laboratories, Mountain View, CA) by using NotI. This EGFP-NotI fragment was first ligated to the 3'-UTR fragment of the His2AvD in pGEM-T (digested with NotI) and then ligated to the exon 4 fragment (digested with HindIII/BglII). The resulting 1.73-kb fragment, which has two SalI sites at the two ends, was first digested with SalI and then ligated to the pONIXAvDGFP-S65T plasmid with exon 4 and GFP-S65T removed by XhoI digestion. The resulting pONIXAvD-EGFP plasmid was verified by using different restriction enzymes and then used to make transgenic flies using standard transformation protocols.

This stronger His-GFP did not allow us to study cycles earlier than cycle 5, presumably because more His-GFP is loaded into the egg, which increases His-GFP in the cytoplasm as well as that incorporated into the chromatin of these embryos. With each division, more His-GFP is incorporated into the chromatin and depleted from the cytoplasm, increasing the contrast. The major advantages of this His2AvD-EGFP transgenic line are that for cycles after cycle 6, reduced laser energy can be used to identify cell-cycle phase progression. Furthermore, we noted that nuclei in imaginal discs are much brighter.

In vivo imaging:

Two-photon laser scanning microscopy (TPLSM) live imaging was described previously (JI et al. 2004). Room temperature was 21.9° ± 1.0° and all the recordings were completed within 20 days. This is very important because room temperatures showed variations between summer and winter months.

Aphidicolin injection:

Aphidicolin (Sigma, St. Louis) was dissolved in 1% DMSO/injection buffer (5 mM KCl, 0.1 mM NaH₂PO₄, pH 7.5) at a concentration of 0.295 mM (RAFF and GLOVER 1988). For the live-imaging experiments with aph. reported here, we modified the protocol. We used shorter egg collections (15 min) to obtain more precise staging. For dechorionation we decreased the bleach treatment (100%) to 1.25 min and increased the distilled water and PBS–Tx (0.1% Triton X) embryo wash to 2–3 min. These modifications are important for recording normal development following injections. Embryos were then aligned on 2.5% agar and transferred to a slide with double-stick tape (3M). Depending on the humidity in the room, embryos were desiccated for 15–17 min and then covered in Halocarbon oil (Halocarbon Inc.). They were injected at 60 (cycle 6), 90 (cycle 10), 120 (cycle 12), or 135 (cycle 13) min after egg deposition (AED). We injected 40–60 embryos within 5 min of each time point. Needles were inserted into the posterior end and 250- to 370-pl injections were made into the center of the embryo. All embryos were kept for 30 min in a humid chamber at 25° AED. Before fixation the oil was washed off with heptane and embryos were then transferred to vials containing 4% formaldehyde in PBS:heptane (1:1) and fixed, with shaking, for 20 min. The formaldehyde layer was then removed and replaced with PBS. Embryos were then hand devitellinized and collected in 1.5-ml tubes. Antibodies to histone H1 and phosphorylated histone H3 were used.

P-element imprecise excision:

The coding region of CG9273 (dRPA2) consists of four exons encompassing 1170 nucleotides in region 38F6 on the second chromosome. CG9273^KG00759 (Bloomington) contains a P{SUPor-P} transposable element in the third exon, rendering it homozygous lethal. To create additional alleles, 2–3 transposase was used to excise this P element (ROBERTSON et al. 1988). Two excision events were null alleles as confirmed by PCR and direct sequencing.

A genetic screen for modifiers of a phenotype caused by excess CycB:

To generate a sensitized genetic background that can be used to identify dominant modifiers, we increased the number of cycB gene copies in the mother to six. Elevated levels of CycB cause multiple nonlethal phenotypes in developing embryos, one of which is a delay in arrival of nuclei to the cortex (Figure 1B). In the vast majority of two cycB embryos (97%), nuclei reach the cortex synchronously within 2 min into interphase of cycle 10 (FOE and ALBERTS 1983; JI et al. 2002). Of six cycB embryos, however, only 30% exhibit such coordination at cycle 10 (Figure 1, B and E, and JI et al. 2002, 2005). Upon arrival to the periphery, nuclei have normal morphology in both two cycB and six cycB embryos.

View larger version (75K): In this window In a new window Download PPT slide   FIGURE 1.— Cycle 10 (A–E) and cycle 14 (A'–E') phenotypes. Embryos were fixed and stained with anti-histone H1 at cycle 10 (A–E). Embryos were also fixed at cycle 14 and stained with anti-histone H1 to label nuclei (red) and antiphosphorylated histone H3 (green), which labels nuclei undergoing mitosis (A'–E'). two cycB (wild-type) embryos at cycle 10 (A) and cycle 14 (A') are shown. Six CycB embryos at cycle 10 (B) and cycle 14 (B') are shown. Arrows in B and E indicate nuclei, which have not reached the cortex at cycle 10. Note the patch of anti-phospho-histone H3, indicating nuclei in M-phase (B' and D'); inset in B'' shows chromosomal bridges. An example of a class I suppressing Df at cycle 10 (C) and cycle 14 (C') is shown. (D and D') A case of a class II suppressing gene (Arp87C). (E and E') An example of a class III suppressing Df. Pole cells (p) are in M-phase at cycle 14 in both six cycB (B') and class II suppressing embryos (D').

 
During cycles 10–14 we find that in two cycB embryos, nuclear morphology remained normal in almost all blastoderm cycles and in all embryos (86–97%, n = 329). Interestingly, six cycB embryos at cycle 10 are normal (91%, n = 56), while in later blastoderm cycles, the frequency of six cycB embryos with normal nuclear morphology drops to <76% (n = 41) and is lowest at cycle 12 (58%, n = 36). Abnormal embryos exhibit one or more of the following: patches of extra mitoses, areas in the blastoderm devoid of nuclei, chromosomal bridges, or macro/micro nuclei (Figure 1B''). In addition, we observe PH3-positive pole cells in six cycB embryos, indicating M-phase at cycle 14 (Figure 1, B' and D'). Despite these abnormalities, hatching rates and development to adults are >90% in both six cycB and two cycB embryos.

In the loss-of-function Df screen previously reported (JI et al. 2002), 70% of the euchromatic genome was analyzed for modifiers of these nonlethal phenotypes. While the elucidation of interactions among cell-cycle components that resulted from this screen (JI et al. 2002, 2005) advanced our understanding of cell-cycle regulation, the screen itself had several potential shortcomings: The genetic background was not isogenic, Df breakpoints were cytologically defined, and many deficiencies were very large. We decided to revisit the screen by using 104 selected deficiencies from the Exelixis and DrosDel collections and 11 deficiencies from the Bloomington collection to fill in gaps from the initial screen (PARKS et al. 2004; RYDER et al. 2004). In addition, new deficiencies were selected to define many of the breakpoints from the Bloomington Df kit. With the new deficiencies, we increased coverage of the euchromatic genome to 83% (Figure 2).

View larger version (34K): In this window In a new window Download PPT slide   FIGURE 2.— Cytogenetic map of polytene salivary gland chromosomes with deficiencies indicating modification of the six cyclin B phenotype. The map includes data from our previous screen (open bars, JI et al. 2002) and from the extended screen reported here (solid bars). The Drosophila euchromatic genome consists of four chromosomes, of which three are represented here with each of the five chromosomal arms (X, 2L, 2R, 3L, and 3R) subdivided into 20 salivary gland units (100 units total). Relative lengths of deficiencies tested for modification are shown. Black bars indicate deficiencies that neither enhanced nor suppressed the six cycB phenotype. Red bars are deficiencies that enhanced and blue bars are those that suppressed the six cycB phenotype. Specific genes that were identified as modifiers are colored similarly; dRPA2 (CG9273) was a suppressor identified in this screen. x indicates a previous falsely identified enhancer that after retesting with a molecularly defined Df in region 61 failed to enhance the six cycB phenotype. dRPA1 Df (84F6) did not enhance or suppress the six cycB phenotype. For specific breakpoints of the deficiencies tested here see the APPENDIX.

 

View this table: In this window In a new window   APPENDIX All deficiency lines used for the screen and their hatching rates in both two cycB and six cycB backgrounds

 
By crossing chromosomal deficiencies into the sensitized background, we found several that modified the six cycB phenotypes at cycle 10, 14, or both: enhancers exacerbated these abnormalities, while suppressors mitigated them and exhibited a phenotype closer to that of two cycB (Figure 1, A and A'). Of the 115 deficiencies tested here, 13 were classified as new suppressors and 6 additional enhancers were found.

As before, we grouped suppressing deficiencies and mutants into one of three categories: Class I suppressors normalized the six cycB phenotype at cycles 10 and 14. Class II suppressed the cycle 10 phenotype, but not the cycle 14 phenotype, while class III (Figure 1 and supplemental material at http://www.genetics.org/supplemental/) improved the cycle 14 phenotype, but not the cycle 10 phenotype (JI et al. 2002). Of the 13 suppressing deficiencies identified here, 4 were class I while 9 were class III. No class new II suppressors were found here.

Identification of dRPA2 as a suppressor of the six cycB phenotypes:

Genomic region 38–39 is interesting because multiple previously described deficiencies enhance the six cycB phenotype (Figure 2). However, within this region we also had previously identified diaphanous as a suppressor (JI et al. 2002). Through candidate gene testing we found another class III suppressor, a P-element insertion into CG9273 (CG9273^KG00759, Figure 2). The identification of two suppressor genes in the same region that is uncovered by several enhancer Df lines suggests that there are one or several strong enhancer genes that are dominant over the suppressive effect of either dia or CG9273^KG00759. Such an enhancer gene(s) in the 38–39 region remains to be identified.

The P element inserts into CG9273 and is a recessive lethal that arrests development in the second larval instar. We mobilized the P element and generated 10 precise and 2 imprecise excisions. We tested 2 precise excision lines. One had 80% and the other 86% normal blastoderms, which were not different from six cycB embryos.

We used imprecise excision of the P element in CG9273^KG00759, here named replication protein A, subunit 2 (dRPA2¹), due to its homology to a human gene (51% amino acid sequence identity) of that name, and identified two new alleles, dRPA2² and dRPA2³. Sequencing of the two excision sites revealed that both excision events deleted part of the second exon plus the two exons downstream of the original P-element insertion site (2L: 20913399–20913400). The imprecise excisions deleted 952 bp of genomic DNA in dRPA2² and removed 1911 bp in dRPA2³. These observations document that the suppressing phenotype of the three dRPA2 alleles is a result of a mutation in the dRPA2 gene. We observed that all three homozygous dRPA2 mutant alleles had hatching rates not different from that of two cycB [wild-type embryos (+95%)]. All animals arrested in the second instar, with a normal larval cuticle pattern. Neither of the two new alleles complemented the original P insertion, nor each other, and over a Df they also arrested in the second instar. Thus all three alleles are amorphic.

Analysis of fixed material revealed that in embryos from mothers heterozygous for any of the three dRPA2 alleles in a six cycB background, >88% demonstrated a normal cycle 14 phenotype (dRPA2¹, 97%, n = 251; dRPA2², 89%, n = 192; dRPA2³, 93%, n = 110). Fewer than 32% of embryos with each allele showed a normal cycle 10 phenotype (dRPA2¹, 24%, n = 136; dRPA2², 23%, n = 30; dRPA2³, 31%, n = 16). All three amorphic alleles produced indistinguishable phenotypes. Thus, dRPA2 was classified as a class III suppressor. We performed experiments with all three alleles, observed no differences between them, and present here data from the ones of which we had the most cases. It should be noted that one small Df (38F6), Df(2L)Exel7080, which uncovered dRPA2, is a class III suppressor as well [25% normal embryos at cycle 10 (n = 56), 93% normal embryos at cycle 14 (n = 194), Figure 2 and APPENDIX].

Previous work studied suppressor genes that began to function before cycle 10 and suppressed the six cycB phenotype at both cycle 10 and cycle 14 (class I) and suggested multiple mechanisms for the phenotypic rescue (JI et al. 2002, 2005). Since class III suppressors mitigated the phenotype at cycle 14 by acting during the blastoderm stages but not beforehand, this provided us with a unique opportunity to uncover differences in these cycles. Before we addressed how lowering of dRPA2 gene copy number rescued the six cycB blastoderm phenotype, we needed to describe differences of blastoderm cycles in embryos with different amounts of CycB.

Blastoderm cycle durations were affected by doses of maternally loaded Cyclin B:

Previous observations of blastoderm cell-cycle times with differential interference contrast (DIC) microscopy established approximate durations of cycles 10, 11, 12, and 13 to be 9, 10, 12, and 21 min, respectively. In addition, cell-cycle phases were estimated by calculating the proportion of fixed embryos in a given phase as a part of the number of embryos examined (FOE and ALBERTS 1983; STIFFLER et al. 1999). Direct observation with two-photon laser scanning microscopy (TPLSM) offers live measurements and therefore provided a more direct and accurate determination of cell-cycle-phase durations (YU et al. 2000; JI et al. 2004). In addition to these analyses of two cycB (wild-type) embryos, further observations established effects of varied CycB levels on the blastoderm cycles to include premature or delayed arrival of nuclei to the cortex as well as changes in nuclear doubling time (STIFFLER et al. 1999). However, a detailed analysis and comparison of the different cycle phases in embryos with varied amounts of maternal CycB were thus far missing. These will be of great benefit in both studying normal blastoderm cycles and further determining the role of CycB in coordination of these cycles.

We generated a new EGFP-labeled histone H2AvD transgenic line (His2AvD-EGFP), which is about seven times stronger than the original histone-GFP-S65T (CLARKSON and SAINT 1999). The use of His2AvD-EGFP embryos with TPLSM greatly facilitated our analyses of nuclear morphology, timing of cell-cycle phases, and total cycle durations. Three nuclear morphologies were distinguished with a 10-sec accuracy: interphase, characterized by round nuclei following telophase, indicating nuclear envelope formation; prophase, marked by sudden loss of sharp and round nuclear morphology, indicating nuclear envelope breakdown, which was similarly observed by intrusion of rhodamine-labeled tubulin into nuclei (YU et al. 2000; JI et al. 2004); and anaphase, when chromatid separation was initiated (Figure 3). The elapsed time between these three different morphologies defines interphase, prophase–metaphase, and anaphase–telophase durations. We reported previously on blastoderm cycles 11 and 12 in two cycB embryos and observed that doubling times differed regionally because of differences in interphase. Doublings were fastest at the two poles and slowest in the middle, attributed to lower nuclear densities at the poles (YASUDA 1992; BLANKENSHIP and WIESCHAUS 2001; JI et al. 2004).

View larger version (102K): In this window In a new window Download PPT slide   FIGURE 3.— Images from a single embryo used to define interphase duration (A and A'), prophase–metaphase duration (B and B'), and anaphase–telophase duration (C and C'). Interphase (i) begins 10 sec (A) after the latest telophase configuration (t in C'). Note the pair of teardrop-shaped nuclei (t), 10 sec before earliest interphase (A). Interphase ends 10 sec before earliest prophase (p in B; note loss of round nuclear morphology). Prophase–metaphase duration (B–B') begins when first condensation is observed (p in B) and ends 10 sec before anaphase (a in B'). Metaphase (m) occurs within this duration. Anaphase–telophase duration (C–C') begins when first movement of chromatids is observed (a in C, note the weak signal from sister chromatids below indicated with a dotted line) and ends in late telophase (t in C').

 
To compare cell-cycle progression of the blastoderm cycles in one cycB, two cycB, and six cycB, we focused on an area between 60% egg length (EL) and 75% EL (100% EL being the anterior pole). We followed nuclei at the four corners of this area. We selected this area because within it, local densities varied less than at the poles. In each region, multiple nuclei were followed for the extent of the blastoderm cycles and their cell-cycle phase durations were averaged. Thus, each embryo provided data from dorsal and ventral cortices, none of which were significantly different.

We found that overall cell-cycle duration steadily increased throughout the blastoderm cycles with all three genotypes and that in all cycles, higher levels of CycB correlated with shorter cell-cycle duration (Figure 4A). These results deviated from previous analyses of these cycles and documented that TPLSM was more accurate in defining the cell-cycle length than DIC microscopy (STIFFLER et al. 1999). Specifically, with TPLSM we observed that one cycB embryos consistently had the longest doubling time and six cycB the shortest, a trend not previously discerned (Figure 4A and STIFFLER et al. 1999).

View larger version (20K): In this window In a new window Download PPT slide   FIGURE 4.— Cell-cycle (A) and cell-cycle phase (B and C) durations during the syncytial blastoderm divisions with varying doses of Cyclin B and the suppression of six cycB by dRPA2. Cell-cycle and phase durations were defined using a Hist2AvD-EGFP tag and two-photon live imaging. For each genotype, five to seven embryos were recorded and analyzed. In all four cycles, one cycB embryos had longer cycles (A) and interphases (B), but shorter prophase–metaphase durations (C) than two cycB embryos, while six cycB embryos had the shortest cycles (A) and interphase durations (B), but longest prophase–metaphase durations (C). Anaphase–telophase durations showed no differences among the genotypes (C). Reducing dRPA2 gene copy number extended interphase more in later blastoderm cycles (B). Looking at total cell-cycle time (A) no differences were observed compared to six cycB embryos. Arrows in B indicate no statistical differences from six cycB (red) or two cycB (black). [P 0.002, in all cases, on the basis of the isotonic regression test (GAINES and RICE 1990).] *Cell-cycle phases presented in seconds. Significant difference from two cycB (based on Tukey–Kramer means comparison test, P < 0.001; SOKAL and ROHLF 1981).

 
TPLSM defined more precisely both the doubling time and cell-cycle phases in the different genotypes. Previously we showed that interphase extension occurred after cycle 7 in two cycB embryos and this extension continued into the blastoderm cycles. With varying doses of CycB this initiation of interphase extension occurred one cycle earlier in one cycB and two cycles later in embryos with more CycB when compared to two cycB embryos (JI et al. 2004). As we show here, these interphase extensions continue during blastoderm cycles. Thus at cycle 10, interphase in embryos with less CycB was longer, whereas in embryos with more CycB interphase was shorter (Figure 4B). Conversely, prophase–metaphase duration was extended with increasing doses of CycB (Figure 4C). Unlike interphase extension, prophase–metaphase extension occurred only after cycle 10. Thus, CycB doses had opposite effects on interphase and prophase–metaphase durations at each blastoderm cycle: Increased CycB had shorter interphase and longer prophase–metaphase durations. Therefore, the differences in total cell-cycle length between one, two, and six cycB embryos were less dramatic than those seen by comparing just interphase duration (Figure 4).

In addition, with TPLSM we observed that anaphase–telophase durations in contrast to other phases remained constant throughout the cycles and in embryos with different CycB doses (Figure 4C).

Thus we saw changes in the maternally regulated cell cycle that commenced well before cycle 10 and continued through cycle 13 and others that did not begin until the start of the blastoderm cycles, such as prophase–metaphase extension.

To address how and whether reducing the number of dRPA2 gene copies affected the six cycB cell cycle and its phases, we analyzed blastoderm cycles of dRPA2²/six cycB embryos. Using the Tukey–Kramer means comparison test (SOKAL and ROHLF 1981) we found that in progressively later blastoderm cycles, the cell-cycle phases became more similar to those of two cycB embryos. Interphase durations in both cycles 12 and 13 were significantly longer than those in six cycB embryos alone and by cycle 13 interphase was comparable to that of two cycB embryos (Figure 4B). A similar trend was observed with prometaphase durations: At cycle 10, these embryos were more similar to six cycB, while at cycle 13 they were more similar to two cycB (Figure 4C). Thus the normalization of the interphase and prophase–metaphase duration in dRPA2²/six cycB embryos occurred gradually and depended on the dose of dRPA2. No such effect was seen on anaphase–telophase duration. While it is well established that Cdk1–CycB is a major player in cell-cycle control, we showed that dRPA2 is also involved in this process in Drosophila. Since our data showed that dRPA2 affects interphase in six cycB embryos it seemed plausible that such an effect could be seen with two cycB embryos. However, we found that reducing gene copy number of dRPA2 just in a two cycB background did not affect total cell-cycle time or its phases (data not shown). We concluded that a dominant function of dRPA2 can be shown only in six cycB embryos.

Onset of a detectable DNA replication effect:

Chk1 homologs delay entry into mitosis by inhibiting cdc25 phosphatases that are required to dephosphorylate and activate Cdk1–CycB (FURNARI et al. 1997). Interphase extension, which is synonymous with delaying mitosis in S–M cycles, requires Grp/Chk1 in blastoderm cycles (JI et al. 2004). In contrast, interphase extension prior to cycle 10 does not require Grp (JI et al. 2004). This suggests the possibility that Grp-dependent regulation of mitosis is absent or inactive prior to cycle 10. To test this idea directly, we injected aph., the most commonly used pharmacological inhibitor of DNA polymerase- and therefore of DNA replication (IKEGAMI 1978), into embryos at various times. Many investigations are published that use aph. extensively and make conclusions about Chk1 activity. However, such experiments do not directly measure Chk1 activity. Therefore, we use the term checkpoint effect to monitor metaphase delay, not checkpoint activity, as has been used by others (for example, SIBON et al. 2000).

We used four different aph. injection time points (60, 90, 120, and 135 min AED) and three different methods to detect a DNA replication effect. With histological analysis, embryos were fixed 30 min after injection, which was enough time to complete several divisions in controls. Normal nuclear morphology served as an indicator that no M-phase was induced. DIC imaging was used to determine whether and when total cell-cycle duration increased. These two protocols would not test specifically for Grp/Chk1 function, but will test generally for the ability to regulate the cell cycle in the presence of aph. Finally, TPLSM analysis was used to determine whether and when interphase was lengthened. This analysis specifically tests for the ability to delay the entry into M-phase in the presence of incomplete DNA replication, which has been shown previously to depend on Grp (YU et al. 2000). Emphasis was placed upon fixed material, however, because many embryos are processed under the same conditions as well as at different time points and with different genotypes. Furthermore, present data can be compared with previously published data using two cycB embryos (RAFF and GLOVER 1988; DEBEC et al. 1996; YU et al. 2000).

Histological analysis of checkpoint effect:

We injected staged one, two, and six cycB embryos with aph. at 60 min AED (cycle 6) and incubated the embryos for 30 min. Failure of checkpoint regulation was scored by chromosomal abnormalities, such as bridges and large and small nuclei, as well as abnormally condensed nuclei, all of which indicate that M-phase was induced before DNA replication was completed (RAFF and GLOVER 1988; DEBEC et al. 1996; YU et al. 2000). Normal interphase or prophase nuclear morphology scored 30 min after injection indicated a delay of M-phase, implicating some checkpoint effect. After injection of aph. at 60 min AED in two cycB embryos, we found very few with normal nuclear morphology, indicating checkpoint failure with this protocol (Figure 5A, <8%). Control injections (1% DMSO) were made with each genotype at 90, 120, and 135 min AED, for a total of 15 control injection series (n > 90 embryos for each control). In Figure 5A we summarize these injections into three control columns. Each column represents the average of the five genotypes injected with 1% DMSO at 90, 120, or 135 min AED (at 90 min, 81–88%, mean = 84%; at 120 min, 85–88%, mean = 86%; at 135 min, 93–98%, mean = 96%, n = 1539).

View larger version (58K): In this window In a new window Download PPT slide   FIGURE 5.— Blockage of DNA replication tests for checkpoint effect in embryos with different amounts of CycB and dRPA2. (A) Control and experimental embryos (x-axis) were injected at 60, 90, 120, or 135 min AED (z-axis). Control embryos were injected with 1% DMSO, experimental embryos with 0.1% aph. solution in 1% DMSO, and both were fixed 30 min later. Embryos were scored as normal (y-axis) if they contained no nuclei with bridges or other division defects; abnormal embryos had at least five nuclei with defects described in the text. *Significant differences in one cycB and six cycB when compared to aph.-injected two cycB embryos. Significant differences in dRPA21 embryos compared to their respective cycB controls (P 0.001). (B) Recordings of single embryos using His2AvD-EGFP and TPLSM microscopy. One (blue), two (black), and six (red) cycB embryos were injected during late prophase of cycles 9 or 11. Recording started (0 min) in early interphase (inter) of cycles 10 or 12, respectively, and until the last moment of interphase. To illustrate checkpoint activity images from recordings are presented here at 5-min intervals within each interphase and ending with the last interphase (inter). Representative cases from each genotype were chosen and their actual interphase duration times are presented here.

 
With no difference and no checkpoint effect detected at the 60-min injection point, we focused our experiments on the later three time points (90, 120, and 135 min AED). Previous work had found that injections around the two latest time points in two cycB embryos resulted in M-phase delays, demonstrating checkpoint effect at these stages (FOGARTY et al. 1997; SIBON et al. 1997; YU et al. 2000). Therefore, we expected that a majority of two cycB embryos at these two latest time points should have normal nuclei and thus could serve as a standard to test for active checkpoint in one and six cycB embryos. We injected aph. into two cycB embryos, first at 90 min AED (cycle 10), when checkpoint activity has not been previously reported, and found that only 25% (n = 158) had normal interphase and prophase nuclear morphology, indicating some, but weak, checkpoint effect. Injection at 120 min AED (cycle 12) in two cycB showed 48% (n = 166) of embryos with normal nuclei, indicating significant checkpoint efficacy and supporting previous work. At the latest time point, 135 min AED (cycles 13), we found that the number of normal embryos was not different from control injections (86%), reflecting complete checkpoint efficacy (Figure 5A, black columns). In summary, we found that in two cycB embryos checkpoint effect can be detected only after 90 min and its efficacy increased over the next 45 min of development.

Checkpoint effect detected by conventional DIC live imaging:

Because results from fixed material could not document any checkpoint effect at 60 min AED, we injected aph. into early prophase of cycles 9, 11, and 12. Prophase was chosen for the following reasons: (1) It can be defined with an accuracy of 20 sec and (2), at this phase, DNA replication was completed. In all embryos we tested for a longer cycle time (cycles 10, 12, and 13) in the cycle following injection, indicating checkpoint effect. We found a significant extension of cycle 12 in two cycB embryos compared to controls (aph.-injected embryos, 30.7 ± 3.2 min, n = 5 embryos; control-injected embryos, 13.5 ± 1.8 min, n = 10). However, at cycle 10 no difference was observed (aph. injection, 10.7 ± 1.5 min, n = 10 embryos; control injections, 9.7 ± 1.0 min, n = 10). Together with the data from the histological analysis, this firmly documented that there is no detectable checkpoint effect prior to cycle 11.

Checkpoint effect defined by TPLSM using histone-EGFP:

While the data reported in the previous sections defined when we can document checkpoint effect, they do not specifically test for an interphase extension or arrest indicating Grp/Chk1 effect. We defined the end of interphase as the time, with 10-sec accuracy, when nuclei lost their round shape. Again, we injected aph. into embryos when all nuclei were in prophase of the cycle preceding the test cycle. With two cycB embryos, we found no interphase extension in cycle 10 (5.5, 4.8, and 5. 7 min, mean = 5.3 ± 0.5 min; Figure 5B), but an extension was observed in cycle 12 (21.7, 25.3, 15.8, and 35.0 min, mean = 24.5 ± 8.1 min). A similar experiment revealed interphase extension in cycle 11 previously (SIBON et al. 1997; YU et al. 2000). Collectively, these data demonstrate the onset of a detectable DNA replication effect beginning with cycle 11 in two cycB embryos.

A link between CycB dose and the onset of a detectable DNA replication checkpoint:

We observed that interphase duration at cycle 11 in six cycB embryos resembled interphase duration at cycle 10 in two cycB embryos (Figure 4B). In other words, interphase duration in embryos with more CycB corresponded to that of earlier cycles in embryos with less CycB. Similar observations were made in grp mutant embryos (SIBON et al. 1997; JI et al. 2004). Thus, a longer interphase depends on both Grp function and the availability of Cdk1–CycB. That is, high Cdk1–CycB levels may override inhibitory effects of Chk1. To address this idea, we compared the time of onset of a detectable checkpoint effect among embryos with different doses of cycB.

Among one cycB embryos injected with aph. at 90 min AED, we observed significantly more embryos with normal nuclear morphology than in two cycB embryos. Injection at 120 min gave even more embryos with normal nuclear morphology, and injection at 135 min AED (cycles 13) produced embryos with normal nuclei in frequencies not different from control levels (Figure 5A, blue columns). This indicates that checkpoint efficacy increased in one cycB embryos with similar dynamics as in two cycB embryos, but its effect was initiated one cycle earlier.

We repeated this experiment with six cycB embryos. We injected aph. at 90 min AED and saw significantly fewer embryos with normal nuclear morphology than in two cycB embryos. Similarly, at the120-min AED injection point, fewer six cycB embryos had normal nuclear morphology compared to two cycB embryos, and even at the latest injection time point (135 min AED) six cycB embryos had only 64% of embryos with normal nuclei (Figure 5, red column). Thus, six cycB embryos never reached checkpoint efficacy observed in one and two cycB and the control embryos (96%). In summary, we found a gradual increase in checkpoint efficacy in six cycB embryos but later initiation than in two cycB. This clearly demonstrates that checkpoint effect can be repressed with high Cdk1–CycB activity.

To further test this, we injected aph. late in the cycle and performed live imaging to ensure that the DNA was replicated. We then used DIC to define the total cell-cycle length of the next cycle. The disadvantages of live imaging are that only one embryo can be handled at a time, and conditions such as temperature vary from day to day and even from one embryo to the next.

All one cycB embryos injected with aph. showed a longer cycle 10 (24.3 ± 5.6 min, n = 8) than after control injections (11.6 ± 1.0, n = 5) while none of the six cycB embryos had this effect (aph.-injected embryos, 9.92 ± 1.1 min, n = 6; control-injected embryos, 10.0 ± 0.63 min, n = 6). Such lengthening did not occur in six cycB embryos until cycle 13 (aph.-injected embryos, 36.4 ± 6.3 min, n = 6; control-injected embryos, 15.9 ± 1.2 min, n = 6). These data indicated that checkpoint effect was observed in cycle 10 for one cycB embryos, in cycle 12 for two cycB embryos, and in cycle 13 for six cycB embryos. This supports the conclusions made from fixed material, namely that checkpoint effect in one cycB is observed earlier and in six cycB later than in two cycB embryos.

Live analysis by TPLSM using histone-GFP supported the above conclusions: As in the previous experiment, we injected aph. into one cycB embryos where all nuclei were in prophase of the cycle preceding the test cycle. Cycle 10 interphase durations were recorded in two embryos (17.6 and 15.3 min, mean = 16.5 min; Figure 5B). These values were significantly longer than those in two cycB embryos and indicated DNA replication checkpoint effect. In six cycB embryos, such an interphase extension was not observed in cycle 12 with two embryos (7.2 and 9.3 min, mean = 8.3 min; Figure 5B). Thus in one cycB embryos we observed interphase extension at cycle 10 and no such extension was seen in two cycB embryos. However, at cycle 12, a clear interphase extension was observed with two cycB embryos but not with six cycB embryos. This directly demonstrated that DNA replication checkpoint effect was manifested in one cycB embryos at cycle 10 and later, at cycle 12 in two cycB embryos, a time point when still no Grp/Chk1 effect was observed with six cycB embryos.

We used these embryos also to address the question of whether aph. affects prophase–metaphase duration. The variation of the data is too broad to make a firm conclusion. In some cases, we observed that chromosome condensation took longer, which was observed by others (YU et al. 2000). This could mean that either a leaky block of Chk1 could affect other checkpoint activities or several checkpoints were initiated at these blastoderm cycles (YU et al. 2000).

dRPA2 suppressed the later onset of Chk1 effect:

The RPA (replication protein A) complex is composed of three different subunits, conserved in yeast, Xenopus, and humans. This complex is involved in DNA replication and DNA damage repair, as it binds and stabilizes ssDNA as the DNA is unwound (WOLD 1997). More interestingly, it was found that Cdk1–CycB phosphorylates RPA2 to disassociate the complex from chromatin (FANG and NEWPORT 1993). In Drosophila it is known that dRPA2 exists in two different phosphorylation states: unphosphorylated during late anaphase to telophase and phosphorylated from midprophase to early anaphase (MITSIS 1995). In yeast, Xenopus extracts, and human cells it has been shown that the RPA complex regulates Chk1 via ataxia–telangiectasia Rad3-related kinase (ATR) (LONGHESE et al. 1996; KIM and BRILL 2001; COSTANZO et al. 2003; ZOU and ELLEDGE 2003). ATR can initiate the phosphorylation of Chk1, which leads to a cell-cycle arrest (LIU et al. 2000). Considering the role of RPA2 in Chk1 function and that Chk1 inhibits Cdk1–CycB via Cdc25 led us to test whether both dRPA2 and CycB are linked to Grp/Chk1 function in the early embryo (FURNARI et al. 1997; ZENG et al. 1998).

To test whether suppression of the six cycB phenotype by lowering dRPA2 gene doses was due to a change in timing of the initiation of checkpoint effect, we injected aph. into dRPA2¹/six cycB embryos. We observed that at 90-, 120-, and 135-min AED injection time points dRPA2¹/six cycB embryos had significantly more cases with normal nuclei than the six cycB embryos. This is interesting because at the 135-min AED injection time point, dRPA2¹/six cycB embryos have more normal nuclear morphology than do six cycB embryos and they reach levels seen in similarly treated two cycB embryos (Figure 5A, pink column).

The most parsimonious interpretation of our data is that reduction of one dRPA2 gene copy somehow allowed the Grp-dependent DNA replication checkpoint to override higher Cdk1–CycB activity and produce a normal cycle 13.

After this observation, it was obvious to test whether reducing dRPA2 in two cycB embryos caused an earlier checkpoint effect. We injected dRPA2¹/two cycB embryos with aph. at 90, 120, and 135 min AED. At 90 min, dRPA2¹/two cycB embryos had significantly more cases with normal nuclear morphology than two cycB embryos but fewer than one cycB embryos, suggesting that dRPA2 potentiated the checkpoint activation that overrode Cdk1–CycB levels. At the 120- and 135-min injection points dRPA2¹/two cycB embryos were not different from two cycB embryos, suggesting that the checkpoint efficacy was close to its full capacity and was dominant over Cdk1–CycB levels (Figure 5).

Since there are three dRPA proteins, we tested whether reducing dRPA1 gene copy number showed a similar effect to reducing dRPA2. We used the deficiency [Df(3R)Exel6147], which uncovered dRPA1, but were unable to document suppression of the six cycB phenotype. It is possible that dRPA1 is not limiting in our system or that dRPA2 has a special role in cell-cycle regulation because Cdk1–CycB directly and specifically phosphorylates dRPA2.

Differences in histone H3 dephosphorylation between one, two, and six cycB embryos:

Cyclin A destruction is necessary for sister chromatid separation (SIGRIST et al. 1995). SU et al. (1999) reported that in embryos from grp– mothers, levels of Cyclin A were higher than in controls and premature loss of the mitosis-specific phosphorylated histone H3 (PH3) was observed in early anaphases. To exit M-phase, Cyclins A, B, and B3 have to be destroyed (SU et al. 1998; PARRY and O'FARRELL 2001). Looking at one, two, and six cycB embryos at cycle 10 and monitoring the disappearance of PH3 we observed late-anaphase loss of PH3 in one and two cycB embryos (Figure 6, A and B). However, in six cycB embryos, like in grp embryos (SU et al. 1999), PH3 loss is observed in early anaphase (Figure 6C). Also like in six cycB, grp embryos showed abnormal nuclei after cycle 12 (SU et al. 1999 and data not shown). Thus, grp and six cycB embryos have similar phenotypes. Reducing dRPA2 gene copy number in six cycB embryos suppresses this phenotype at cycle 10 (Figure 6D).

View larger version (31K): In this window In a new window Download PPT slide   FIGURE 6.— Histone H3 dephosphorylation begins earlier in anaphase of a six cycB embryo than in that of one, two cycB, and dRPA2/six cycB embryos. Late anaphase of one cycB (A, A'), two cycB (B, B'), and dRPA2/six cycB (D, D') and early anaphase of six cycB (C, C') embryos fixed at cycle 10 and stained with an antibody to histone H1 (red) and the phosphorylated form of histone H3 (green) are shown. Note the earlier loss of PH3 in six cycB nuclei (C, C') in 6 of 10 cases observed as compared to the later anaphase of one cycB in 4 of 4 cases, of two cycB (B, B') in 5 of 5 cases, and of dRPA2/six cycB (D, D') in 4 of 4 cases. This indicates that six cycB embryos have delayed Grp activity when compared with two cycB (SU et al. 1999).

 

Maternal regulation of the blastoderm cycles and onset of the DNA replication checkpoint:

It has been proposed (model I) that the ability of the DNA replication checkpoint to regulate the entry into M-phase is not active before cycle 11 in wild-type (two cycB) embryos. This model is based on observations that the earliest defects in interphase duration can be detected only after cycle 10 in grp (dChk1) or mei-41 (dATR) mutant embryos (SIBON et al. 1997, 1999). In this view, before cycle 10 in two cycB embryos, nuclei have normal morphology because the DNA replication machinery is abundant. The exponentially increasing numbers of nuclei, however, titrate components of this machinery to a critical level after cycle 10, thereby inducing the checkpoint activation, interphase extension, and delay of M-phase in a Grp- and Mei-41-dependent manner (BLUMENTHAL et al. 1974; SIBON et al. 1997, 1999).

We propose an alternative model (II): Checkpoint function is active all the time, but before cycle 10 checkpoint activity is too low and is overridden by the high level of maternal Cdk1–CycB. The difference between the two models is that in model I, checkpoint is activated by a critical amount of the replication machinery. In model II, at a critical concentration of Cdk1–CycB, this kinase (Cdk1) can no longer override checkpoint activity. Several observations are not compatible with model I, but are with model II. First, here we show that checkpoint activity does not depend on a specific number of nuclei, number of rounds of divisions, or time after fertilization, but on the amount of CycB–Cdk1. It occurs earlier in one cycB embryos with fewer nuclei and later in six cycB embryos with more nuclei.

Second, model I does not explain interphase extension before cycle 10 and the slight increase of interphase even in grp mutant embryos after cycle 10 (SIBON et al. 1997; YU et al. 2000; JI et al. 2004).

Third, Grp protein is required for the degradation of cyclin A in the presence of cycloheximide as early as cycle 4 (SU et al. 1999). Although the effect of Grp on CycA degradation may be different from its effect on replication checkpoint activation, this observation suggests that Grp is present and functional before cycle 10. How then do nuclei enter M-phase at the normal time when aph. is applied before cycle 10, i.e., show no DNA replication checkpoint? We reason that successful execution of the checkpoint requires the inhibitory effect of Grp to overcome the M-phase-promoting effect of Cdk1–CycB. Early embryos might have too few nuclei, thus limited numbers of replication forks, to trigger sufficient Grp/Chk1 activity necessary to overcome the relatively high levels of Cdk1–CycB. Despite this situation, nuclear morphology is normal because replication machinery is not limited before cycle 10, S-phase is rapidly completed, and nuclei can successfully undergo a normal mitosis.

Fourth, during normal S-phase of either yeast or mammalian cells, a low level of replication checkpoint activity is observed (SHECHTER et al. 2004; PETERMANN et al. 2006). This low level of checkpoint activity can be detected as phosphorylation on Chk1 in S-phase cells without any replication stress or DNA damage (PESCHIAROLI et al. 2006). Physiological regulation of Chk1 is under the control of similar factors of the DNA replication checkpoint machinery, and thus it was proposed that DNA replication per se generates lesions that activated the checkpoint pathway (SORENSON et al. 2004). Alternatively, but not mutually exclusively, this constitutively low level of replication checkpoint activation may be due to transient signaling from the replication forks, which does not lead to cell cycle arrest, but serves as a mechanism to coordinate the firing of replication origins, thereby moderating the rate of S-phase (SHECHTER et al. 2004).

Model II not only accounts for our observation with the aph. experiments, but also accounts for observation of the CycB effects on the replication checkpoint effect. In one cycB embryos the Grp/Chk1 effect is observed earlier, presumably because levels of Cdk1–CycB become limiting earlier, whereas in six cycB embryos, these occur later. Using PH3 staining on anaphase chromosomes as a measurement for Chk1-dependent Cyclin A degradation indicates that Chk1 is not functioning in six cycB embryos before cycle 11, possibly because it is overridden by the abundance of Cdk1–CycB in these embryos (SU et al. 1998). We propose that in six cycB embryos at cycle 11 or later, Cdk1–CycB activity is still too high and forces the nuclei into mitosis at a time when the DNA replication machinery is limited, resulting in precocious M-phase and abnormal nuclei.

Reduced dRPA2 gene copy number suppresses the six cycB phenotype:

Here, we address how dRPA2 might suppress the six cycB phenotype at cycle 14. dRPA2 is a subunit of a highly conserved heterotrimeric complex of proteins that make up the RPA complex. All three subunits contain DNA-binding domains, which stabilize ssDNA as it is unwound at the replication fork (WOLD 1997). This stabilized DNA allows for Cdc45 and DNA polymerase- to initiate DNA replication (FAIRMAN and STILLMAN 1988; WOLD and KELLY 1988). Additional roles for RPA have been implicated in DNA repair and recombination (COVERLEY et al. 1991, 1992; MOORE et al. 1991; ZOU and ELLEDGE 2003).

As proposed above, in six cycB embryos, during the blastoderm cycles, elevated levels of Cdk1-CycB override Chk1 and the nuclei divide before DNA replication is completed, leading to abnormal nuclei. We speculate that reducing dRPA2 is likely to slow DNA replication because RPA can facilitate DNA replication by unwinding dsDNA and by modulating the activities of several enzymes, such as DNA helicases, DNA polymerases, and primases (WOLD 1997). This would result in less RPA coated, primed DNA and ssDNA. Such a DNA structure may potentiate the TopBP1-mediated ATR–ATRIP kinase activation, leading to stronger Chk1 activation (ZOU and ELLEDGE 2003; KUMAGAI et al. 2006). Thus in dRPA2/six cycB embryos, a stronger Chk1 activation would have a stronger inhibitory effect on Cdk1–CycB activity that cancels out the effect caused by extra Cdk1–CycB (Figure 4). This interpretation for the suppressive effect of RPA2 on the six cycB phenotype suggests an antagonistic relationship between the DNA replication checkpoint activation and Cdk1–CycB activity in regulating the transition from the preblastoderm cycles to the blastoderm cycles.

Control of early embryonic development:

The earliest divisions in the embryo (with the exception of mammals) are maternally regulated until the zygotic genome takes over. On the basis of many observations with different animals such as Xenopus and Drosophila, a simple concept developed: Early cell cycles are invariant, synchronous, and lack both gap phases until the transition to zygotic control occurs. The time point at which this change happens has been called the midblastula transition (MBT). It is assigned to the specific cycle when zygotic transcription is activated. Furthermore, G2 is induced in Drosophila and G1 and G2 in Xenopus into the abbreviated cell cycles and cell division is patterned (NEWPORT and KIRSCHNER 1982a,b; EDGAR and O'FARRELL 1989; YASUDA and SCHUBIGER 1992; ORR-WEAVER 1994).

This concept is attractive; however, as with many simple concepts, the more we learn the harder it is to accept the concept at face value. For example, synchronous divisions have never been observed in Caenorhabditis elegans. Even in Drosophila, the simplification that early cell cycles are synchronous and equal in length is incorrect as interphase durations steadily increase after cycle 7 and metasynchronous mitoses are observed as early as cycle 4 (JI et al. 2004). These changes occur long before cycle 14, the time that has been designated by many as the MBT. The data presented here clearly demonstrate a change in the maternal program as the embryo develops: A DNA replication checkpoint is first detectable after cycle 10, but becomes increasingly robust in the subsequent cycles, indicating that the ability to regulate M-phase by checkpoints is not completely "off" or "on." In addition, we found that the DNA replication checkpoint is detectable earlier in one cycB and later in six cycB embryos, clearly indicating that changes do not have to occur at a specific stage. A gradual attainment of full checkpoint function is also supported by the fact that aph. injections before cycle 11 can stall/delay the nuclear cycle, but not the centrosomal cycle (RAFF and GLOVER 1988; DEBEC et al. 1996; WAKEFIELD et al. 2000; B. EDGAR, personal communication; and our unpublished observations). These results are not compatible with the idea of an invariant maternal program for pre-MBT cycles.

Another simplification is that changes that define the MBT are events that envelope the entire embryo. In sea urchins the deceleration of the cell cycle in the micromeres occurs long before these changes have been observed in the macromeres. Thus an MBT has not been proposed for this animal. But differences are also observed among the blastoderm cycles in Drosophila, where divisions in the middle of the embryo are slower than at the poles (JI et al. 2004), correlating with their nuclear densities (YASUDA 1992; BLANKENSHIP and WIESCHAUS 2001).

We have previously reported that the initiation of the zygotic program also does not occur suddenly from off to on. fushi tarazu (ftz) transcripts are first observed at cycle 9 in one or the other embryo and in one or the other nucleus. Transcription gradually increases over the next three cycles (cycles 9–12). This gradual increase is a consequence of the dose-dependent repressor Tramtrack (TTK), where with one gene dose of maternal ttk, initial transcription of ftz occurs one cycle earlier, and conversely extra copies of ttk result in initial transcription of ftz one cycle later (PRITCHARD and SCHUBIGER 1996). We interpreted these data as a decline of TTK during cycles 8–10 to a threshold level where TTK repression is insufficient, enabling low-level transcription of ftz (PRITCHARD and SCHUBIGER 1996).

Despite the many observations that do not fit the MBT concept, textbooks, reviews, research articles, and grant proposals still hold on to it, hindering progress in the understanding of how maternally controlled development declines and terminates and zygotic programming gradually takes over.

We acknowledge fly stocks from Bloomington and plasmids from Rob Saint. T.T. Su and M. Schubiger provided us with extensive help and criticism of this manuscript. We also thank Fajun Yang and Lee Zou for helpful discussions. This work was supported by a grant from the National Science Foundation to G.S.

BLANKENSHIP, J. T., and E. WIESCHAUS, 2001 Two new roles for the Drosophila AP patterning system in early morphogenesis. Development 128: 5129–5138.[Medline]

BLUMENTHAL, A. B., H. J. KRIEGSTEIN and D. S. HOGNESS, 1974 The units of DNA replication in Drosophila melanogaster chromosomes. Cold Spring Harbor Symp. Quant. Biol. 38: 205–223.[Abstract/Free Full Text]

BRODSKY, M. H., J. J. SEKELSKY, G. TSANG, R. S. HAWLEY and G. M. RUBIN, 2000 mus304 encodes a novel DNA damage checkpoint protein required during Drosophila development. Genes Dev. 14: 666–678.[Abstract/Free Full Text]

CLARKSON, M., and R. SAINT, 1999 A His2AvDGFP fusion gene complements a lethal His2AvD mutant allele and provides an in vivo marker for Drosophila chromosome behavior. DNA Cell Biol. 18: 457–462.[CrossRef][Medline]

COSTANZO, V., D. SHECHTER, P. J. LUPARDUS, K. A. CIMPRICH, M. GOTTESMAN et al., 2003 An ATR- and Cdc7-dependent DNA damage checkpoint that inhibits initiation of DNA replication. Mol. Cell 11: 203–213.[CrossRef][Medline]

COVERLEY, D., M. K. KENNY, M. MUNN, W. D. RUPP, D. P. LANE et al., 1991 Requirement for the replication protein SSB in human DNA excision repair. Nature 349: 538–541.[CrossRef][Medline]

COVERLEY, D., M. K. KENNY, D. P. LANE and R. D. WOOD, 1992 A role for the human single-stranded DNA binding protein HSSB/RPA in an early stage of nucleotide excision repair. Nucleic Acids Res. 20: 3873–3880.[Abstract/Free Full Text]

DEBEC, A., R. F. KALPIN, D. R. DAILY, P. D. MCCALLUM, W. F. ROTHWELL et al., 1996 Live analysis of free centrosomes in normal and aphidicolin-treated Drosophila embryos. J. Cell Biol. 134: 103–115.[Abstract/Free Full Text]

EDGAR, B., 1995 Diversification of cell cycle controls in developing embryos. Curr. Opin. Cell Biol. 7: 815–824.[CrossRef][Medline]

EDGAR, B. A., and P. H. O'FARRELL, 1989 Genetic control of cell division patterns in the Drosophila embryo. Cell 57: 177–187.[CrossRef][Medline]

EDGAR, B. A., F. SPRENGER, R. J. DURONIO, P. LEOPOLD and P. H. O'FARRELL, 1994 Distinct molecular mechanisms regulate cell cycle timing at successive stages of Drosophila embryogenesis. Genes Dev. 8: 440–452.[Abstract/Free Full Text]

FAIRMAN, M. P., and B. STILLMAN, 1988 Cellular factors required for multiple stages of SV40 DNA replication in vitro. EMBO J. 7: 1211–1218.[Medline]

FANG, F., and J. W. NEWPORT, 1993 Distinct roles of cdk2 and cdc2 in RP-A phosphorylation during the cell cycle. J. Cell Sci. 106: 983–994.[Abstract]

FOE, V. E., and B. M. ALBERTS, 1983 Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis. J. Cell Sci. 61: 31–70.[Abstract]

FOGARTY, P., S. D. CAMPBELL, R. ABU-SHUMAYS, B. S. PHALLE, K. R. YU et al., 1997 The Drosophila grapes gene is related to checkpoint gene chk1/rad27 and is required for late syncytial division fidelity. Curr. Biol. 7: 418–426.[CrossRef][Medline]

FURNARI, B., N. RHIND and P. RUSSELL, 1997 Cdc25 mitotic inducer targeted by chk1 DNA damage checkpoint kinase. Science 277: 1495–1497.[Abstract/Free Full Text]

GAINES, S. D., and W. R. RICE, 1990 Analysis of biological data when there are ordered expectations. Am. Nat. 135: 310–317.[CrossRef]

HUANG, J., and J. W. RAFF, 1999 The disappearance of cyclin B at the end of mitosis is regulated spatially in Drosophila cells. EMBO J. 18: 2184–2195.[CrossRef][Medline]

IKEGAMI, S., T. TAGUCHI, M. OHASHI, M. OGURO, H. NAGANO et al., 1978 Aphidicolin prevents mitotic cell division by interfering with the activity of DNA polymerase-alpha. Nature 275: 458–460.[CrossRef][Medline]

JACOBS, H. W., J. A. KNOBLICH and C. F. LEHNER, 1998 Drosophila Cyclin B3 is required for female fertility and is dispensable for mitosis like Cyclin B. Genes Dev. 12: 3741–3751.[Abstract/Free Full Text]

JI, J. Y., M. HAGHNIA, C. TRUSTY, L. S. GOLDSTEIN and G. SCHUBIGER, 2002 A genetic screen for suppressors and enhancers of the Drosophila cdk1-cyclin B identifies maternal factors that regulate microtubule and microfilament stability. Genetics 162: 1179–1195.[Abstract/Free Full Text]

JI, J. Y., J. M. SQUIRRELL and G. SCHUBIGER, 2004 Both cyclin B levels and DNA-replication checkpoint control the early embryonic mitoses in Drosophila. Development 131: 401–411.[Abstract/Free Full Text]

JI, J. Y., J. CREST and G. SCHUBIGER, 2005 Genetic interactions between Cdk1-CyclinB and the Separase complex in Drosophila. Development 132: 1875–1884.[Abstract/Free Full Text]

KIM, H. S., and S. J. BRILL, 2001 Rfc4 interacts with Rpa1 and is required for both DNA replication and DNA damage checkpoints in Saccharomyces cerevisiae. Mol. Cell. Biol. 21: 3725–3737.[Abstract/Free Full Text]

KUMAGAI, A., J. LEE, H. Y. YOO and W. G. DUNPHY, 2006 TopBP1 activates the ATR-ATRIP complex. Cell 124: 943–955.[CrossRef][Medline]

LIU, Q., S. GUNTUKU, X. S. CUI, S. MATSUOKA, D. CORTEZ et al., 2000 Chk1 is an essential kinase that is regulated by Atr and required for the G(2)/M DNA damage checkpoint. Genes Dev. 14: 1448–1459.[Abstract/Free Full Text]

LONGHESE, M. P., H. NEECKE, V. PACIOTTI, G. LUCCHINI and P. PLEVANI, 1996 The 70 kDa subunit of replication protein A is required for the G1/S and intra-S DNA damage checkpoints in budding yeast. Nucleic Acids Res. 24: 3533–3537.[Abstract/Free Full Text]

MITSIS, P. G., 1995 Phosphorylation and localization of replication protein A during oogenesis and early embryogenesis of Drosophila melanogaster. Dev Biol. 170: 445–456.[CrossRef][Medline]

MOORE, S. P., L. ERDILE, T. KELLY and R. FISHEL, 1991 The human homologous pairing protein HPP-1 is specifically stimulated by the cognate single-stranded binding protein hRP-A. Proc. Natl. Acad. Sci. USA 88: 9067–9071.[Abstract/Free Full Text]

MURRAY, A. W., and T. HUNT, 1993 The Cell Cycle: An Introduction. W. H. Freeman, New York.

NEWPORT, J., and M. KIRSCHNER, 1982a A major developmental transition in early Xenopus embryos: I. Characterization and timing of cellular changes at the midblastula stage. Cell 30: 675–686.[CrossRef][Medline]

NEWPORT, J., and M. KIRSCHNER, 1982b A major developmental transition in early Xenopus embryos: II. Control of the onset of transcription. Cell 30: 687–696.[CrossRef][Medline]

ORR-WEAVER, T. L., 1994 Developmental modification of the Drosophila cell cycle. Trends Genet. 10: 321–327.[CrossRef][Medline]

PARKS, A. L., K. R. COOK, M. BELVIN, N. A. DOMPE, R. FAWCETT et al., 2004 Systematic generation of high-resolution deletion coverage of the Drosophila melanogaster genome. Nat. Genet. 36: 288–292.[CrossRef][Medline]

PARRY, D. H., and P. H. O'FARRELL, 2001 The schedule of destruction of three mitotic cyclins can dictate the timing of events during exit from mitosis. Curr. Biol. 11: 671–683.[CrossRef][Medline]

PESCHIAROLI, A., N. V. DORRELLO, D. GUARDAVACCARO, M. VENERE, T. HALAZONETIS et al., 2006 SCFbetaTrCP-mediated degradation of Claspin regulates recovery from the DNA replication checkpoint response. Mol. Cell 23: 319–329.[CrossRef][Medline]

PETERMANN, E., A. MAYA-MENDOZA, G. ZACHOS, D. A. GILLESPIE, D. A. JACKSON et al., 2006 Chk1 requirement for high global rates of replication fork progression during normal vertebrate S phase. Mol. Cell. Biol. 26: 3319–3326.[Abstract/Free Full Text]

PRITCHARD, D. K., and G. SCHUBIGER, 1996 Activation of transcription in Drosophila embryos is a gradual process mediated by the nucleocytoplasmic ratio. Genes Dev. 10: 1131–1142.[Abstract/Free Full Text]

RAFF, J. W., and D. M. GLOVER, 1988 Nuclear and cytoplasmic mitotic cycles continue in Drosophila embryos in which DNA synthesis is inhibited with aphidicolin. J. Cell Biol. 107: 2009–2019.[Abstract/Free Full Text]

ROBERTSON, H. M., C. R. PRESTON, R. W. PHILLIS, D. M. JOHNSON-SCHLITZ, W. K. BENZ et al., 1988 A stable genomic source of P-element transposase in Drosophila melanogaster. Genetics 118: 461–470.[Abstract/Free Full Text]

RYDER, E., F. BLOWS, M. ASHBURNER, R. BAUTISTA-LLACER, D. COULSON et al., 2004 The DrosDel collection: a set of P-element insertions for generating custom chromosomal aberrations in Drosophila melanogaster. Genetics 167: 797–813.[Abstract/Free Full Text]

SHECHTER, D., V. COSTANZO and J. GAUTIER, 2004 ATR and ATM regulate the timing of DNA replication origin firing. Nat. Cell Biol. 6: 648–655.