Argonaute-1 functions as a mitotic regulator by controlling Cyclin B during Drosophila early embryogenesis

Abstract

The role of Ago-1 in microRNA (miRNA) biogenesis has been thoroughly studied, but little is known about its involvement in mitotic cell cycle progression. In this study, we established evidence of the regulatory role of Ago-1 in cell cycle control in association with the G₂/M cyclin, cyclin B. Immunostaining of early embryos revealed that the maternal effect gene Ago-1 is essential for proper chromosome segregation, mitotic cell division, and spindle fiber assembly during early embryonic development. Ago-1 mutation resulted in the up-regulation of cyclin B-Cdk1 activity and down-regulation of p53, grp, mei-41, and wee1. The increased expression of cyclin B in Ago-1 mutants caused less stable microtubules and probably does not produce enough force to push the nuclei to the cortex, resulting in a decreased number of pole cells. The role of cyclin B in mitotic defects was further confirmed by suppressing the defects in the presence of one mutant copy of cyclin B. We identified involvement of 2 novel embryonic miRNAs—miR-981 and miR-317—for spatiotemporal regulation of cyclin B. In summary, our results demonstrate that the haploinsufficiency of maternal Ago-1 disrupts mitotic chromosome segregation and spindle fiber assembly via miRNA-guided control during early embryogenesis in Drosophila. The increased expression of cyclin B-Cdk1 and decreased activity of the Cdk1 inhibitor and cell cycle checkpoint proteins (mei-41 and grp) in Ago-1 mutant embryos allow the nuclei to enter into mitosis prematurely, even before completion of DNA replication. Thus, our results have established a novel role of Ago-1 as a regulator of the cell cycle.—Pushpavalli, S. N. C. V. L., Sarkar, A., Bag, I., Hunt C. R., Ramaiah, M. J., Pandita, T. K., Bhadra, U., Pal-Bhadra, M. Argonaute-1 functions as a mitotic regulator by controlling Cyclin B during Drosophila early embryogenesis.

Coordination between the timing and onset of biological events is critical for cell division and differentiation during development. The first 13 syncytial nuclear divisions in Drosophila are maternally controlled and mainly consist of the S and M phases, with short or undetectable gap phases (1). The absence of cytokinesis, along with shortened G₁ and G₂ phases, provides an ideal platform for screening new factors that participate in cell proliferation and for analyzing sister chromatid segregation and mitotic defects.

Drosophila Ago-1 (dAgo1), a homologue of Arabidopsis Ago-1, was initially identified as a regulator of the wingless signaling pathway (2). Ago-1 is involved predominantly in post-transcriptional gene silencing, but affects many biological functions, such as oocyte and stem cell fate determination (3, 4), DNA and histone methylation in plants, RNA interference (RNAi) in animals (5), quelling in fungi (6), heterochromatin formation and maintenance in fission yeast and Drosophila (7, 8), and chromatin organization, especially in Arabidopsis, Schizosaccharomyces pombe, and Drosophila (5, 8, 9). Recent studies in S. pombe and Trypanosoma brucei clearly demonstrated the role of 2 core components of the RNAi machinery in mitotic cell division: Ago-1 and Dcr-1 (10, 11). Deletion of the Ago-1 or Dcr-1 genes in S. pombe is characterized by centromere dysfunction, chromosomal abnormalities that are seen as lagging chromosomes in the late anaphase, and chromosome loss and telomeric defects during mitosis or meiosis (11). In addition, Ago-1 is critical for cell cycle checkpoints in fission yeast, as the amino terminus directly binds to 14-3-3 proteins and induces a cell cycle delay at the G₂/M boundary (12). Ago-2, another member of the same gene subfamily, is necessary for proper nuclear migration, pole cell formation, and cellularization during early embryonic development in Drosophila (13). Ago-1 colocalizes with Polycomb foci in developing embryos and is involved in nuclear clustering of Pc-G targets for Polycomb-mediated silencing (14). Ago-1 operates in multiple pathways, including microRNA (miRNA) Cdk1 biosynthesis (15), heterochromatin assembly (16), cell division, and proliferation, in higher eukaryotes. Furthermore, deletion of the chromosomal region spanning human Ago-1 results in tumor formation (17, 18).

Embryonic lethality and Ago-1-dependent cell cycle progression in mutant embryos suggest that Ago-1 controls genomic stability and cell proliferation in Drosophila. In Drosophila, zygotic transcription has a negligible role in the nuclear divisions that occur during early embryogenesis. As the embryos reach the syncytial stage (10–13), the cell cycle is progressively slowed to account for complete elongation of longer transcripts and also for repair of any DNA damage that occurs in the course of development. However, mutations in the checkpoint genes such as mei-41 and grp (Chk1) do not show any delay during the late cleavage cycles (19, 20).

The spatiotemporal regulation of cyclin-dependent kinases and cyclins controls the entry into and exit from mitosis. Cdk1 and CycB form an active heterodimer that acts as a mitosis-promoting factor (MPF). Moreover, expression of mouse cyclin B1 (Ccnb1) relies on the key factors involved in miRNA biogenesis and function (i.e., Dicer, Drosha, Ago-1, and Ago-2; ref. 21). Cyclin B is also localized on microtubules during the blastoderm stage of Drosophila (22), and transgenic flies carrying 6 copies of maternal CycB gene clearly have shorter microtubules and fewer asters than do wild-type embryos (23). However, proper coordination of cyclins and Cdks, along with cytoskeletal dynamics, is important in cell cycle progression.

In contrast to mammalian p53, Drosophila p53 induces apoptosis but has no influence in cell cycle arrest during stress (24). Embryos derived from mei-41 or grp mutant females fail to terminate the final syncytial division, are defective in initiating zygotic transcription, and do not cellularize (19, 20, 25).

MATERIALS AND METHODS

Fly stocks

Drosophila Ago-1 y¹w^67c23, P{w[+mC]=lacW}AGO1^k08121/CyO,cn ¹ mutant alleles (hereafter referred to as Ago-1 mutant) were obtained from the Bloomington stock center (Indiana University, Bloomington, IN, USA; http://flystocks.bio.indiana.edu). y w f dicer1Q^1147X/TKG, alleles (hereafter referred to as Dicer-1) were obtained from G. Cavalli (Institute of Human Genetics, Montpellier, France), CycB ²/CyO from W. E. Theurkauf (State University of New York, Stony Brook, NY, USA), Ago-1 transgenic stock (with Ago-1t) from Stephen Cohen (European Molecular Biology Laboratory, Heidelberg, Germany; ref. 26), and multiple balancer stocks from James A. Birchler (University of Missouri, Columbia, MO, USA). Each mutation is described in the Flybase (Indiana University; http://flybase.org), unless otherwise noted. Flies were cultured on standard food medium at 25°C.

Immunohistochemistry

Drosophila embryos were dechorionated with 50% commercial bleach for 2–3 min and fixed with paraformaldehyde/heptane mix for 15–20 min. The embryos were processed as described earlier (16), and antibodies were used in the following dilutions: goat anti-Vasa, 1:20; rat anti-α-tubulin, 1:20;, rabbit anti-PH3, 1:50; and mouse anti-actin, 1:30. Goat- or donkey-raised FITC-, Cy3-, and Cy5-conjugated secondary antibodies were used at a dilution of 1:50. The embryos were mounted in Vectashield medium (Vector Laboratories, Burlingame, CA, USA) containing propidium iodide (PI) or DAPI and viewed by confocal microscope (FV1000; Olympus, Tokyo Japan).

Total protein isolation

Nearly 0.1 g of embryos was collected and homogenized in lysis buffer (6% SDS, 1 mM EDTA, 2 mM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 10 μg/ml pepstatin). The samples were boiled at 95°C for 5 min and centrifuged at 12,000 rpm at 4°C for 10 min. The supernatant was collected, and Western blot analysis was performed according to standard protocols. The following antibodies were used for hybridization of the blots with Ago-1 (1:500; Abcam, Cambridge, UK): cyclin B (1:300), p53 (1:300), and β-actin (1:500) (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Real-time PCR analysis

Total RNA was isolated by Trizol (Invitrogen; Life Technologies, Bangalore, India), treated with DNase (Ambion; Life Technologies), and purified with an RNeasy Mini Kit (Qiagen, New Delhi, India). RNA (3 μg) was reverse transcribed with Superscript II reverse transcriptase (Invitrogen; Life Technologies) and oligodT primer, according to the manufacturer's instructions. Quantitative real-time PCR was performed with the ABI 7900 system (Applied Biosystems; Life Technologies) in triplicate. Reaction mixtures contained 12.5 μl of SYBR Green I dye master mix (Applied Biosystems; Life Technologies), 2 pmol each of forward and reverse primers, and 5 μl of 1:100 cDNA. Thermocycling conditions included initial denaturation at 95°C for 10 min, followed by 40 cycles at 95°C (30 s), 55°C (30 s), and 72°C (30 s). Each experiment was synchronized according to a baseline set automatically by the software. The relative amount of gene expression was compared with the endogenous control (18S rRNA) and calculated with SDS 2.3 software (Applied Biosystems; Life Technologies).

The primer sequences used for PCR were grp (chk1) forward (F), 5′-ACCCGATTCTTTGTGACCAC-3′, and reverse (R), 5′-AACCACCCAGTCTTTCGATG-3′; p53 F, 5′-AATGCCCATCCAACCACTTA-3′, and R, 5′-AAGGCTCAACGCTAAGGTGA-3′; CycB F, 5′-TTCATCACGGACGACACCTA-3′, and R, 5′-CGAGTAGCGTCGAAGGAAGT-3′; mei-41 F, 5′-CCTTCATCCGGCACTTTTTA-3′, and R, 5′-CGCAAAGGTAGGACAGCTTC-3′; 18S rRNA F, 5′-CCTTATGGGACGTGTGCTTT-3′, and R, 5′-CCTGCTGCCTTCCTTAGATG-3′; Ago-1 F, 5′-CGAAAGGTGAACCGTGAGAT-3′, and R, 5′-GTAGTAACCCTCCGGGGAAC-3′; and Wee1 F, 5′-TCAAATGACCTGGTGCACAT-3′, and R, RP-5′-GGATGCATCATCTGGGCTAT-3′.

Design and synthesis of anti-miRNA oligonucleotides (AMOs) and miR mimics

To improve the function of synthetic oligonucleotides, an optimized LNA/DNA-PS AMO targeting miR-981 and miR-317 (Exiqon, Woburn, MA, USA) was microinjected directly into Drosophila embryos 3 times consecutively, leading to inactivation of the miRNAs (27). AMOs with DNA composition probably worked in this context only because of the use of microinjection. Each compound was a 15-mer with 10 LNA and 5–6 DNA nucleotides (28). The AMO and miRNA mimics were microinjected into the dechorionized embryos 0–15 min after egg laying (AEL) at a concentration of 25–200 nM.

miR mimics were miR-981, 5′-TTCGTTGTCGACGAAACCTGCA-3′; miR-317, 5′-TGAACACAGCTGGTGGTATCCAGT-3′; miR-AMO, miR-317 G*A*A*C*A*C*A*G C*T*G*G*T*T*A*T (LNA/DNA); and miR-981 T*C*G*T*T*G*T**C*G*A*C*G*A*A*A (LNA/DNA).

miRNA expression studies

A bioinformatic search was carried out using the microRNA Targets and Expression program (http://www.microrna.org/microrna/home.do) to identify putative miRNAs regulating the CycB gene. Total RNA from early embryos (syncytial cycle 12–13) of control and Ago-1 mutations were isolated with Trizol, and RNase-Free DNase treatment was used to remove DNA contaminants. cDNA was prepared with miScript (Qiagen), according to the manufacturer's instructions. Quantitative RT-PCR was performed with the primers for miR-317 and miR-981. 2S rRNA was used as an internal control. The primer sequences were miR-981, 5′-TTCGTTGTCGACGAAACCTGCA-3′; miR-317, 5′-TGAACACAGCTGGTGGTATCCAGT-3′; and 2S rRNA, 5′-TACAACCCTCAACCATATGTAGTCCAAGCA-3′.

Cell culture and luciferase assay

S2 cells were grown at 25°C in Schneider's Drosophila medium, supplemented with l-glutamine. The 3′-UTR of CycB containing the binding site for miR-981 was amplified from the cDNA of early embryos. The amplicon was isolated, purified, and cloned in a psi-CHECK-2 dual luciferase vector by using the NotI and XhoI cloning sites (Promega, Madison, WI, USA). The target site for miR-981 was mutated by using the Quick Change Site-Directed Mutagenesis kit (Stratagene; Agilent Technologies, Palo Alto, CA, USA). S2 cells (2×10⁶) were cotransfected (Invitrogen; Life Technologies) with 500 ng wild-type or mutated CycB 3′-UTR and anti-miR 981 (5 μl of 10 μM stock; ThermoFisher Scientific, Waltham, MA, USA), with SiLentFect used as the transfection reagent. 2′-O-Methyl and 3′-cholesterol modifications were incorporated, with the addition of 5 arbitrary bases to both the 5′ (UCUUA) and 3′ (ACCUU) ends of the anti-miR sequence, to increase transfection efficiency. After 72 h of transfection, the lysates obtained were subjected to a dual luciferase reporter assay (Promega). The Renilla luciferase values were normalized to the firefly luciferase values (Promega). The relative normalized values were measured from 3 independent experiments. Primers for CycB 3′-UTR amplification are F, 5′-TCCAAGGCGGACTGGAAG-3′, and R, 5′-CCACTCTCTCTCTCGGTGTCC-3′.

The sequence of synthesized anti-miR-981 is 5′-mU.mC.mU.mU.mA.mG.mC.mC.mC.mA.mA.mA.mG.mC.mA.mA.mU.mC.mG.mU.mC.mG.mC.mC.mC.mG.mA.mA.mC.mC.mU.mU-3′-Chl (ThermoFisher Scientific).

RESULTS

Haploinsufficiency of Ago-1 leads to mitotic defects

Ago-1 mutation in Drosophila is recessive lethal in the late embryonic stage. Ago-1 protein is maternally deposited, and zygotic expression starts close to stage 9, at around 2.5 h AEL (2). Real time-PCR experiments showed that the Ago-1^k08121/CyO mutant (P-element insertion) and the Ago-1 excisions Ago-1Ex28/CyO and Ago-1 Ex37/CyO had considerable reduction of transcripts in 0–1, 1–2, 2–3, 3–4 h embryos compared to yw^67c23 (Fig. 1A). Embryos from the Ago-1^k08121/CyO mothers (haploinsufficiency of maternal gene product), Ago-1 excision alleles generated by the imprecise P-element excision (Ago-1^Ex-28/CyO, Ago-1^Ex-37/CyO), were collected (16). The yw^67c23 embryos and embryos from Ago-1 (Ago-1^rev/CyO) revertant stock generated by the precise P-element excision were used as the control. Early embryos (0–2 h) were dechorionated, processed, and stained with PI. During the early syncytial nuclear divisions, Ago-1 mutant (Ago-1^k08121/CyO) embryos exhibited several mitotic defects: frequent mitotic catastrophic events (Fig. 1Ba), with fragmented chromosomes that appeared as disorganized, scrambled masses of chromatin; mitotic chromosome breaks; abnormal mitotic figures (Fig. 1Bb); and chromatin bridges (Fig. 1Bc), typical features of lagging chromosomes on the metaphase plane. In a few (15–25%) cases, rounding of anaphase nuclei led to a telophase-like configuration before separation was observed (Fig. 1Bd). The survival rate of embryos from Ago1[k08121]/CyO mothers is only 50%, because CyO/CyO and Ago1[k08121]/Ago1[k08121] embryos do not survive.

Figure 1.

Loss of Drosophila Ago-1 causes chromosomal defects during mitosis in early embryos. A) Real-time PCR analysis showing the reduction of maternal Ago1 transcript at various time intervals during early embryogenesis. B) Embryos (0–3 h) from yw^67c23and Ago-1^K08121, Ago-1^Ex-28, Ago-1^Ex-37, and Ago-1^Rev (revertant, precise excision of the P element) were fixed and stained with PI to visualize chromosomal defects. The yw^67c23 and Ago-1^Rev were used as controls. a) Syncytial blastoderm embryos showing nuclei in yw^67c23, Ago-1 mutations (Ago-1^K08121, Ago-1^Ex28, Ago-1^Ex37), and Ago-1^Rev. In the Ago-1 mutation, the nuclei in the early embryos appear as large, fragmented masses of chromatin. Number of nuclei in the syncytial blastoderm stage in the mutants is very much reduced compared to that in the controls. b) Sister chromatid separation was affected in Ago-1 mutant embryos, giving rise to abnormal mitotic figures (inset). c) Enlarged view of wild-type embryos showing a regular and complete separation of the chromosomes in the anaphase stage, whereas lagging chromosomes and chromatid bridges were seen at the same anaphase stage in the Ago-1 mutant embryos (inset). d) Ago-1 mutant embryos exhibited rounding of the daughter nuclei, as in the telophase stage, although they were still in the early anaphase (inset). e) A large number of fallout nuclei are seen in the Ago-1 mutant embryos, thus lessening nuclear density in the syncytial blastoderm embryos. C) Quantitative depiction of fallout nuclei in yw^67c23, Ago-1 mutants, and Ago-1^Rev embryos. Compared to control embryos of y^w67c23 and Ago-1^Rev, the mutants show a large number of fallout nuclei, indicating the presence of abnormal nuclei that were removed. D) The Ago-1 mutants showed a high frequency of chromosomal defects in the anaphase. Scale bars = 10 μm.

We next analyzed nuclear fallout in syncytial blastoderm embryos (nuclear cycles 9–13), a process that normally removes abnormal nuclei by internalizing them into the yolk and digesting them. The number of fallout nuclei located 2–20 μm below the cortex in the syncytial blastoderm stage was scored (Fig. 1Be; ref. 29). The yw^67c23 embryos or embryos carrying revertant Ago-1 (Ago-1^rev/CyO) alleles were found to contain 2.5% fallout nuclei, whereas Ago-1 mutant embryos had 59% (Fig. 1Be, C). Our data indicated that nearly 75–80% of Ago-1 mutant embryos displayed multiple mitotic defects relative to the control (Fig. 1D). Next, to understand whether mitotic defects are related to Ago-1 mutation (Ago-1^k08121/CyO), 1 copy of the Ago-1 transgene (driven by tubulin promoter) was introduced into the Ago-1 heterozygous (Ago-1^k08121/CyO) mutant embryos. Introduction of the Ago-1 transgene prevented the embryonic lethality associated with the Ago-1 (Ago-1^k08121/CyO) mutation, whereas in 60–70% of the embryos, the mitotic defects were abolished, and a normal phenotype was restored (Supplemental Fig. S1 and ref. 26). (Ago-1/CyO females were crossed with the Ago-1 transgenic males, and the Ago-1/Ago-1 transgenic females were used to collect the early embryos to study the mitotic phenotypes.) To examine whether the balancer chromosome has influence on Ago-1-related mitotic defects, the Ago-1^k08121/CyO and Ago-1^k08121/+ genotypes were compared. Similar mitotic defects were observed in both the embryos, indicating that the CyO balancer chromosome had no influence on the embryonic phenotype. To verify nuclear fallout and visualize the existence of free centrosomes in the fallout nuclei yw^67c23 and Ago-1 (Ago-1^k08121/CyO), we immunostained mutant embryos with anti-Centrosomin antibody. Free centrosomes in the absence of chromosomes were visible in the cortex of Ago-1^k08121/CyO, but not in control yw^67c23 embryos (Fig. 2), indicating the loss of abnormal nuclei from the cortex.

Figure 2.

Immunostaining of yw^67c23 and Ago-1^K08121 embryos with Centrosomin antibodies. Free centrosomes without the chromosomes are seen in the early embryos, indicating the loss of nuclei from the cortex or the presence of fallout nuclei in the Ago-1 mutation. Scale bars = 5 μm.

Ago-1 is required for mitotic synchrony

To ascertain whether mitotic synchrony is disrupted by the Ago-1 mutation, we immunostained precortical embryos (nuclear cycles 9–11) of the yw^67c23 and Ago-1 (Ago-1^k08121/CyO) mutants with anti-phospho-histone H3 (PH3) antibody, a known marker for identification of cells undergoing mitosis, to examine histone H3 phosphorylation on Ser10. In yw^67c23 embryos, H3Ser10 phosphorylation was detected simultaneously in all the nuclei when chromosome condensation occurred with the onset of mitosis. The precortical (Fig. 3A) and syncytial blastoderm (Fig. 3B) embryos carrying the Ago-1 mutation showed unevenly spaced and asynchronous nuclei, whereas nuclei from the control embryos were evenly spaced and underwent synchronous division.

Figure 3.

Mitotic asynchrony in Drosophila Ago-1^K08121 embryos. Embryos were immunostained with anti-PH3 (a mitosis marker) antibody (Ab, e and Bb, e). DNA was counterstained with PI (Aa, d and Ba, d). Merged images are included (Ac, f and Bc, f). A) Staining of the precortical embryos of yw^67c23 (a–c) and Ago-1 (d–f) with anti-PH3 Ser10 antibody revealed the presence of unevenly spaced, asynchronously dividing nuclei in the Ago-1 mutants. B) Staining of syncytial blastoderm embryos of yw^67c23 revealed that all the chromosomes are in mitosis and are stained with the PH3 antibody, indicating mitotic synchrony (a–c), whereas in the Ago-1 mutant embryos, both mitotically dividing and nondividing nuclei are present, indicating a lack of mitotic synchrony (d–f). Scale bars = 10 μm.

Ago-1 is required for cytoskeleton organization and pole cell formation

Cytoskeletal proteins mediate the processes of nuclear division and nuclear migration (30). Embryos of the yw^67c23 and Ago-1 (Ago-1^k08121/CyO) mutant were immunostained with β-actin and α-tubulin antibodies to study the changes in the organization of the cytoskeleton. In yw^67c23 embryos, actin-based pseudocleavage furrows completely encircled the metaphase chromosomes, forming a honeycomb-like structure (Fig. 4A). In contrast, large gaps surrounding the chromosomes or groups of chromosomes with incompletely surrounded actin filaments were found in the Ago-1 mutant embryos. A dense thickening of the actin filaments surrounding groups of chromosomes was also observed. The presence of lagging chromosomes and chromatin bridges suggests deformities in the organization of the spindle apparatus in the Ago-1 (Ago-1^k08121/CyO) mutants.

Figure 4.

Defective cytoskeleton organization and nuclear migration in the Ago-1^K08121 embryos. A) Early stage yw^67c23 (a–c, g–i) and Ago-1^K08121 (d–f, j–l) embryos were immunostained with β-actin antibody (FITC; b, e, h, k) and PI (DNA dye; a, d, g, j). Merge images are included (c, f, i, l). a–f) In control yw^67c23 embryos, the actin cytoskeleton formed a regular honeycomb-like structure (a–c; ×20) that was disrupted by the Ago-1 mutation (d–f). g–i) Higher-magnification images from the yw^67c23 embryos in which the chromosomes are completely surrounded by pseudocleavage furrows. j–l) In the Ago-1 mutation, groups of nuclei were incompletely surrounded by the actin cytoskeleton. B) Embryos from yw^67c23 (a–c) and Ago-1^K08121/CyO (d–f) were stained with α-tubulin (red; b, e) antibody and DAPI (blue; a, d) DNA stain. In the control embryos, the microtubules were aligned perpendicularly to the chromosomes at the equatorial plane, as shown in the merged images (red and blue; c, f). In the case of Ago-1 mutants, the microtubule alignment on the equatorial plane was disrupted and bound all along the chromosomes, leading to the formation of the lagging chromosomes. C) Relative percentage of mitotic spindle defects in control yw^67c23 and Ago-1^K08121/CyO embryos. D) Syncytial blastoderm embryos were stained with anti-Vasa (green), a polar granule component that selectively stains pole cells. There are relatively fewer pole cells in Ago-1 embryos. E) Average number of pole cells in yw^67c23 and Ago-1^K08121/CyO mutant embryos. Scale bars = 10 μm.

Microtubule polymerization and depolymerization during chromosome movements play a crucial role in cell cycle progression (31). In the case of the yw^67c23 embryos, the chromosomes were properly aligned along the metaphase plane and were perpendicularly attached to the α-tubulin fibers, whereas this organization was disrupted in the Ago-1 mutant embryos. The mutant embryos showed abnormal mitotic figures (Fig. 4B) with spindle fibers losing selectivity for centromeric sites and attached in a random fashion all along the chromosomes. An average 76.0% of the Ago-1 (Ago-1^k08121/CyO) mutant embryos exhibited spindle fiber defects compared to only 3% in the yw^67c23embryos (Fig. 4C). We further analyzed the nuclear migration events necessary for the formation of pole cells (the precursors of germ cells). In nuclear cycle 9, few nuclei migrated and reached the posterior cortex where they formed pole cells during the 9/10 cycle transition. Since gross defects in spindle fiber assembly were found in the Ago-1 mutant embryos, we anticipated disruption of pole cell formation. The number of pole cells in the yw^67c23 and Ago-1 (Ago-1^k08121/CyO) mutant embryos was determined by immunostaining with an anti-Vasa antibody that stains the pole cell granule component. As expected, the number of pole cells was reduced in most of the Ago-1 (Ago-1^k08121/CyO) mutant embryos tested (Fig. 4D). On average, the number of pole cells was reduced from 22 in the yw^67c23 embryos to 8–9 in the Ago-1 mutant (Ago-1^k08121/CyO) embryos (∼90 embryos were studied; Fig. 4E). Only in rare cases (2% embryos) were the number of pole cells in the Ago-1^k08121/CyO mutants found to be close to the number observed in the yw^67c23 embryos.

Increased expression of Cdk1-cyclin B in Ago-1 mutants

Chromosomal abnormalities, such as mitotic asynchrony, lagging chromosomes, and defective cytoskeleton structure, that are associated with Ago-1 mutation suggest defective cell cycle checkpoint regulation. The early embryonic cycle is regulated notably by CycA, CycB, and CycB3 and the late syncytial cycles by mei-41 and grp. Early syncytial divisions in Drosophila embryos occur without significant oscillation at the levels of Cyc/Cdk1 activity. However, injection of inhibitor of CycB or stabilized CycB prevents exit from mitosis (32). To address the role of Ago-1 in the regulation of cyclins, we isolated total RNA from embryos of syncytial cycle 12–13 (1.5-2 h AEL) and performed real-time PCR to measure the CycA, -B3, and -B mRNA that controls the cell's commitment to mitosis (33). Of interest, in the Ago-1 (Ago-1^k08121/CyO) mutant embryos, CycB transcripts, as well as protein levels, were significantly up-regulated, whereas those of Cyc A and CycB3 were down-regulated by nearly 2-fold (Fig. 5A, B). PH3 staining acts as a reporter of Cdk1 activity (32). Localized inactivation of Cdk1 results in exit from mitosis, even in the presence of cyclins and Cdk1. Similarly, the loss of PH3 protein localization during the anaphase indicates the loss of Cdk1 and exit from mitosis (34). In the Ago-1 mutant embryos, a strong accumulation of PH3 was observed in the anaphase chromosomes, indicating high Cdk1 activity (Fig. 5C). Further, inhibitors of cyclin-dependent kinases are necessary for proper coordination of cell cycle events. Wee1, a Drosophila Cdk1 inhibitor, phosphorylates the tyrosine 15 residue on Cdk1 and inhibits its activity. Therefore, it regulates mitosis entry during the nuclear cycles of the syncytial blastoderm embryo (35). To address whether Ago-1 regulates Cdk1 by controlling the level of maternal wee1 kinase, we collected yw^67c23 and Ago-1 (Ago-1^k08121/CyO), mutant embryos at 0–1, 1–2, 2–3, 3–4 h AEL. Real-time PCR was performed from total RNA with wee1 primers. Wee1 transcripts were down-regulated in Ago-1 mutant embryos when compared to yw^67c23 embryos (Fig. 5D). Our results suggest that reduction in the levels of maternally deposited wee1 kinase may activate Cdk1 and cause rapid mitosis. Thus, increased Cdk1/CycB activity finally produces shorter microtubules with a decreased metaphase and longer anaphase that eventually leads to delayed exit from mitosis (23).

Figure 5.

Ago-1 mutation affected expression of cell cycle genes. A) Real-time PCR analysis of cell cycle genes showing reduced transcript levels of CycA, CycB3, p53, mei-41, and grp and increased levels of CycB in the Ago-1^K08121/CyO embryos (syncytial cycles 12–13). Data are derived from 3 independent experiments. Bars represent mean ± sd relative RNA expression normalized to an internal control (18S rRNA), which had a minimal effect on the Ago-1 mutant background. B) Western blot analysis depicting reduced levels of p53 protein and increased expression of CycB in the Ago-1^k08121/CyO embryos relative to the β-actin loading control. C) Ago-1^k08121/CyO embryos show PH3 staining in the anaphase stage of early syncytial cycles, which indicates increased activity of Cdk1.D) Real-time PCR analysis showing reduced maternal transcripts of wee1 Drosophila Cdk1 inhibitory kinase at different stages of embryogenesis in the Ago-1 mutant. The relative expression of each transcript is presented, normalized to an internal control (18S rRNA). E) Embryos of maternal genotype Ago-1^k08121/CycB2 showing normal mitosis and actin cytoskeleton compared to abnormal mitosis in the Ago-1^k08121/CyO mutant during early embryogenesis.

Suppression of mitotic defects in the Ago-1 mutants by reduction of cyclin B dosage

To investigate whether the observed mitotic defects are indeed due to increased Cdk1/CycB activity, we performed rescue experiments by reducing the levels of cyclin B in Ago-1^k08121/CyO mutant embryos. Ago-1 mutant females were crossed with CycB2/CyO males, and early embryos from Ago-1^k08121/CycB2 mothers (containing only 1 mutant copy of CycB) were collected and further stained with PI to visualize defects during mitosis. The reduction in dose of maternal CycB completely suppressed nuclear division defects, including mitotic catastrophe, mitotic asynchrony, chromatid bridges, and abnormal mitosis. Staining with β-actin antibody revealed the restoration of defective cytoskeleton to normal, whereas PH3 staining confirmed the restoration of mitotic synchrony (Fig. 5E). The total rescue of the mitotic defects in the early embryos from the mothers of Ago-1^k08121/CycB2 clearly indicates that increased expression of CycB resulted in these defects.

Role of Ago-1 in the expression of proteins involved in the DNA replication/damage checkpoint pathway

Drosophila p53 is not necessary for radiation-induced cell cycle arrest (24), and p53 mutation exhibits increased radiosensitivity and genomic instability (36). To estimate the levels of p53 transcripts from early embryos (cycles 12–13), we performed quantitative real-time PCR. A marked reduction of p53 mRNA in the Ago-1 (Ago-1^k08121/CyO) mutant was observed, indicating the role of Ago-1 in cell death or apoptosis (Fig. 5A, B). A large number of mitotic defects associated with Ago-1 mutation may postulate the improper functioning of the checkpoint genes grp and mei-41, components of the well-conserved DNA-replication/damage checkpoint pathway (19, 20). The yw^67c23embryos have a longer interphase that accounts for the repair of DNA replication and DNA damage defects and allows the complete transcription of longer transcripts. The down-regulation of both mei-41 and grp transcripts associated with Ago-1 mutation (Ago-1^k08121/CyO) may have allowed the nuclei to enter into mitosis even before the completion of DNA replication or repair of damaged DNA (Fig. 5A).

Ago-1 and miRNA machinery mediate regulation of cyclin B and p53

RNAi effector enzymes such as Dcr-1 are necessary for cleavage of pre-miRNA to mature miRNA, and cells with mutations in Dcr-1 and Ago-1 fail to process functional miRNAs. To investigate whether p53 and CycB are the immediate targets of Drosophila miRNAs, we estimated the amount of p53 and cyclin B proteins in the early embryos of wild-type, Ago-1, and Dcr-1 mutants. The levels of p53 and cyclin B would remain the same in both Ago-1 and Dcr-1 mutants if these genes are not controlled by the miRNA. If p53 is controlled by Ago-1- and/or Dcr-1-processed miRNAs, up-regulation of p53 protein should occur in the absence of such miRNAs. Instead, down-regulation of p53 occurred in the Dcr-1 and Ago-1 mutants, indicating an indirect miRNA-mediated regulation of p53. The repression of various p53 activators by the Ago-1 mutation could be mediated by miRNA. On the other hand, the levels of cyclin B were strongly elevated in both the mutants (Fig. 6A). Therefore, our preliminary data indicate that Ago-1 may regulate cyclin B through an miRNA-dependent pathway. Since Dcr-1, like Ago-1, is also involved in miRNA biogenesis, we studied the Dcr-1 mutant embryos for mitotic defects. Similar to the Ago-1 mutants, the Dcr-1 mutants displayed mitotic defects, such as chromatid bridges, telophase defects, mitotic catastrophe, abnormal mitosis, and a reduced number of nuclei in the syncytial blastoderm stage, caused by loss of abnormal nuclei (Supplemental Fig. S2).

Figure 6.

Role of miRNAs in Ago-1-mediated control of p53 and CycB. A) Western blot analysis showing the relative expression (proteins/β-actin) of p53 and cyclin B proteins in Ago-1 and Dcr-1 mutant embryos. Loss of 2 miRNA biogenesis factors (Ago-1 and Dcr-1) had a differential effect on p53 and cyclin B expression, which is summarized in the histograms. In Dcr-1 and Ago-1 mutation the levels of p53 were reduced, whereas the CycB level increased dramatically. Data are derived from triplicate blots. B) Predicted base pairing of mature miR-317 and miR-981 seed sequences in the 3′-UTR of CycB (http://www.targetscan.org). C) Expression of miR-317 and miR-981 regulating cyclin B was studied in control and Ago-1^k08121/CyO embryos (syncytial cycles 12–13). We observed decreased levels of miR-317 and miR-981 in the Ago-1^k08121/CyO embryos. Effect was more dramatic in miR-981 expression. D) Expression of miRNAs regulating cyclin B in wild-type and Ago-1^k08121/CyO during embryonic stages 1–4 was determined. We observed significant reductions in the levels of miR-317 and miR-981 during stage 4. E) Quantitative real-time PCR analysis of CycB transcripts of Drosophila Ago-1 mutant embryos microinjected without mimics, with the miR mimic, and with mismatched mimics. Relative fold changes were calculated after normalization with 18s rRNA. Error bars = sd. *P < 0.05.

A bioinformatics search identified miRNAs that regulate CycB based on the microRNA Targets and Expression program (http://www.microrna.org/microrna/home.do). Using this program, we identified 2 miRNAs, miR-317 and miR-981, that have high mirMVR scores. The program identified these 2 miRNAs binding to the 3′ UTR of Drosophila CycB mRNA, indicating that CycB is a target for known Drosophila miRNAs (Fig. 6B). To establish their role, we performed miRNA expression studies from early embryos of the Ago-1 mutant (syncytial cycles 12–13; Fig. 6C). A clear down-regulation of both the miRNAs was observed in the Ago-1 mutant embryos, which resulted in up-regulation of cyclin B. In addition, we performed real-time PCR to study the expression of both miR-317 and miR-981 at stages 1–4 during early embryogenesis. We observed a significant reduction in the expression of both the miRNAs in the Ago-1 mutant only in stage 4 of early embryogenesis (Fig. 6D). Reduced expression of miRNAs regulating cyclin B resulted in a clearcut up-regulation of cyclin B.

miR-981 regulated cyclin B

Earlier analyses have shown that a defective Ago-1 gene has a greater reduction of miR-981 expression than of miR-317 (Fig. 6C). To determine whether miR-981 and miR-317 repress the expression of CycB, we microinjected miR-317 and miR-981 mimics (a double-stranded 23-nt form of miR-981 and miR-317) separately into 2 sets of Ago-1 mutant embryos, which normally express relatively low levels of endogenous, mature miR-981 and miR-317 miRNAs. In both cases, the miR mimic caused a reduction in the level of cyclin B transcript. The level was reduced by only 10–15% of its normal expression in the Ago-1 mutant embryos with the miR-317 mimic, whereas the level of cyclin B reduction was 70–75% with the miR-981 mimic (Fig. 6E). Furthermore, introduction of anti-AMO, which tightly binds and inactivates the miRNA, inhibited the binding of miR-317 and miR-981 to the cyclin B 3′-UTR, leading to up-regulation of cyclin B (Supplemental Fig. S3). Of note, the effect of AMO was more pronounced with miR-981 than with miR-317. These results strongly suggest that miR-981 acts as potential and effective regulator of CycB.

miRNA target validation is performed with reporter gene constructs transfected in cultured Drosophila S2 cells. The UTRs of CycB containing putative target sites for Drosophila miRNAs were separately cloned downstream of the Renilla luciferase open reading frame contained in the psi-CHECK-2 vector. The recombinant plasmids were then transfected in cultured Drosophila embryonic S2 cells, and luciferase activity was measured 48 h after transfection. The psi-CHECK-2 vector without CycB UTR sequences served as the control. The CycB 3′-UTR was targeted by miR-981, thereby showing a marked inhibition of the reporter expression by 50% (2-fold) in S2 cells (Fig. 7), when compared to the vector alone. To confirm the specificity of miR-981 on the CycB UTR, we introduced a mutant version of the CycB 3′-UTR in a reporter plasmid that is complementary to the seed region of miRNA. When mutant construct was transfected in the S2 cells, the reporter activity was found to be significantly higher than that obtained with the wild-type construct, providing strong evidence of a direct interaction between miR-981 and CycB UTR. It should be noted that miR-981 is one of the best scoring miRNAs (mirSVR score=−0.2189) that is predicted to bind to CycB mRNA. Further CycB 3′-UTR constructs were cotransfected along with anti-miR 981 in Drosophila S2 cells. We found that introduction of anti-miR-981 [antisense oligonucleotide (ASO)] in S2 cells containing wild-type CycB 3′-UTR resulted in increased expression of the reporter (luciferase) gene when compared to the 3′-UTR alone. The transfection of the vector construct that has the mutation in the miRNA binding site in the 3′-UTR did not exhibit much change in the presence or absence of anti-miR. These observations have indicated specific effects of miR-981 on the CycB gene that control early mitosis in the rapidly dividing Drosophila embryo.

Figure 7.

miR-981 controls CycB gene expression. Wild-type and mutant 3′-UTR of CycB was cloned into the psi-CHECK2 vector. Drosophila S2 cells were cotransfected with the CycB 3′-UTR and with or without anti-miR 981. Expression of a luciferase reporter gene containing wild-type CycB 3′-UTR with an intact miR-981 binding site was increased by 2-fold when anti-miR 981 was added to Drosophila S2 cells. Mutant CycB3′-UTR, however, did not show any change in expression of reporter gene, with or without anti-miR 981 transfection. Thus, this enhancement in luciferase expression is specific to the miR-981 seed region within the CycB 3′-UTR. CycB 3′-UTR mut represents the UTR in which the binding site for the miRNA was mutated. AM-981 represents the anti-miR 981 used in the cotransfection study. CycB 3′-UTR represents the wild-type UTR with an intact miR-981 binding site.

DISCUSSION

In the present study, we identified the role of Ago-1 in regulating cyclins, Cdk1 inhibitors, and p53 in Drosophila embryos. In the rapidly dividing cells of the Drosophila embryo, Ago-1 mutation led to severe mitotic disruption, as evidenced by chromosome fragmentation, mis-segregation, and abnormal mitosis during the precortical syncytial cycles. The present results demonstrate that Ago-1 modulated developmental arrays associated with establishing the cell cycle control, seeing that Ago-1 mutation down-regulated Cyc A, CycB3, p53, mei-41, and grp, but upregulated CycB transcripts. The reduction in grp and mei-41levels suggests that the replication and DNA damage checkpoints are perturbed, allowing progression of mitosis before completion of DNA replication or DNA repair, which shows that the embryonic lethality is associated with Ago-1 mutation. These results are consistent with earlier findings that, in Drosophila, DNA replication checkpoint genes are activated to delay cell cycle progression during late cleavage stages (37). In the syncytial blastoderm, the essential replication checkpoint function is to prevent DNA damage and ensure proper repair by delaying the cell cycle (37). The reduced mei-41 or grp levels in the Drosophila embryo due to Ago-1 mutation may cause rapid progression from the S phase to mitosis, even before replication is complete.

The syncytial blastoderm stage in Drosophila involves only the S/M cycles and the expression patterns of cell cycle proteins; for example, mitotic cyclins are necessary for entry into and exit from mitosis (32). CycB is localized to microtubules during the blastoderm stage of Drosophila (22), and increased Cdk1/CycB activity causes shorter microtubules with a decreased metaphase and longer anaphase duration that leads to defective mitosis (23). We also observed the effect of miRNAs on CycB: Ago-1 affects the biogenesis of miRNAs that regulate CycB, leading to the increased expression of CycB. The elevated CycB levels found in the Ago-1 mutants showed that the microtubules were less stable and probably did not produce enough force to push the nuclei into the cortex, resulting in the observed decrease in pole cell formation. Thus, Ago-1 is necessary to ensure proper assembly of the mitotic spindle by controlling the timing of CycB expression, a prerequisite for proper nuclear migration during embryonic development. Moreover, less stable microtubules require a longer time to form proper metaphase structures (23, 33). It is a well-established fact that PH3 staining indicates Cdk1 activity (32, 34). In Ago-1 embryos, the PH3 signal often persists over the entire chromosome through the anaphase, whereas it is restricted to the telomeric regions during the wild-type anaphase, indicating the reminiscence of Cdk1 activity. In Drosophila wee1, a Cdk1 inhibitory kinase, functions downstream of mei-41 and is necessary for regulating the activity of Cdk1 (38). Ago-1 mutant embryos reduced maternal wee1 transcript and hence reduced inhibitory phosphorylation of Cdk1, leading to rapid mitosis. Mutants with reduced maternal wee1 cause premature entry into mitosis, spindle fiber defect, and chromosome condensation defect (39).

The embryonic phenotypes such as mitotic asynchrony, mitotic catastrophe, and disruption of the actin cytoskeleton that are associated with Ago-1 mutation were restored to a normal pattern in the presence of 1 copy of mutant CycB, indicating the role of CycB in mitotic progression. From our results, we confirm that Ago-1 is necessary to ensure proper mitotic progression by controlling the timing of Cdk1/CycB expression, a prerequisite for proper microtubule assembly and nuclear migration during embryonic development (23). The cell cycle checkpoint proteins control the timing of the regulatory pathways, such as DNA replication and chromosome segregation, with high fidelity. As in Drosophila, mammalian Atr and Chk1 are essential during embryogenesis (39). One of the reasons for the observed segregation defects in these mutations in Drosophila is that damaged DNA or incompletely replicated DNA fails to trigger metaphase-to-anaphase delay (19, 20). Recent data in mice indicate that depletion in the miRNA processing factors down-regulates a large number of cell cycle genes, including CycB1 (Ccnb1), implying that miRNAs positively regulate cell-cycle genes. In the current study, miRNAs, such as miR-774, miR-1186, and miR-466d-3p, activated CycB1 and regulated the cell cycle (21). To our surprise, we observed that miRNA down-regulated CycB1 during early embryogenesis in Drosophila in the presence of wild-type Ago-1. Our data clearly indicate that Ago-1 functions as a mitotic regulator by spatiotemporal regulation of Cdk1-CycB1, Chk1 (grp), and mei-41.

In Drosophila, p53 has no role in damage-induced cell cycle arrest, but is absolutely necessary for genomic stability, which is achieved by its apoptotic rather than cell cycle function (24). We speculate that decreased levels of p53 in the Ago-1 mutant may be associated with genomic instability in the early embryos when subjected to stress. Both mei-41 and grp function in the same genetic pathway and maternal mei-41 and grp are necessary for wild-type cell cycle delays during the late syncytial blastoderm stage (19, 37) The reduction in maternal mei-41 and grp caused mitotic defects during the later syncytial divisions, indicating that gene expression defects in the late embryos are secondary consequences of the mitotic errors (19).

Recent studies have identified that noncoding miRNAs act as regulators of gene expression in multicellular eukaryotes and have been implicated in various diseases. miRNAs control cell cycle progression by regulating the cyclin-dependent kinases, cyclins, and cyclin-dependent kinase inhibitors (40). Mutation in miRNA-processing factors (Ago-1 and Dcr-1) up-regulate the levels of CycB mRNA and protein, which indicates their involvement in CycB regulation. In this study, we identified the miRNA-dependent regulatory circuit that up-regulates CycB expression. We therefore suggest that expression of miR-981 in Drosophila embryo and its ability to fine tune CycB make it an optimal mechanism for maintaining a balanced level of CycB expression. To date, no mammalian homologue of miR-981 has been identified. The miRNAs miR-981 and miR-317 are also Ago-1-associated miRNAs, with greatly reduced expression under Ago-1 knockdown conditions in S2 cells (41). The in silico prediction of miR-317 in the red flour beetle (insect class) indicates that components of cytoskeleton are its target (42). We found strong homology between Drosophila and the red flour beetle in the miR-317 mature sequence, and we postulate that down-regulation of miR-317 in Drosophila might have affected the normal functioning of the cytoskeleton, as well as CycB, in the Ago-1 mutant embryos.

In the case of mammals, it has been reported that in several tumor cell lines, the level of Ago-1 is significantly lower than in nontumor cells (43). Wilms' tumor exhibits the deletion of a region of human chromosome 1 that harbors the Ago-1 gene and is also associated with neuroectodermal tumors (17, 18). The haploinsufficient maternal Ago-1 mutant, with all its mitotic defects, survives to develop into the adult only if zygotic transcription of Ago-1 occurs at about stage 9, in the absence of which it dies during the late embryonic stage (2).

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.