| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gamete Biology; |
INRA,3 UMR85 Physiologie de la Reproduction et des Comportements, CNRS, UMR6175, Université de Tours, Haras Nationaux, 37380 Nouzilly, France
CNRS UMR6061 Université de Rennes 1,4 Institut de Génétique et Développement de Rennes, 35043 Rennes, France
Cell Cycle Group,5 Electron Microscopy Department, A.N. Belozersky Institute, Moscow State University, 119992 Moscow, Russia
ABSTRACT
Maturation of immature bovine oocytes requires cytoplasmic polyadenylation and synthesis of a number of proteins involved in meiotic progression and metaphase-II arrest. Aurora serine-threonine kinases—localized in centrosomes, chromosomes, and midbody—regulate chromosome segregation and cytokinesis in somatic cells. In frog and mouse oocytes, Aurora A regulates polyadenylation-dependent translation of several mRNAs such as MOS and CCNB1, presumably by phosphorylating CPEB, and Aurora B phosphorylates histone H3 during meiosis. We analyzed the expression of three Aurora kinase genes—AURKA, AURKB, and AURKC—in bovine oocytes during meiosis by reverse transcription followed by quantitative real-time PCR and immunodetection. Aurora A was the most abundant form in oocytes, both at mRNA and protein levels. AURKA protein progressively accumulated in the oocyte cytoplasm during antral follicle growth and in vitro maturation. AURKB associated with metaphase chromosomes. AURKB, AURKC, and Thr-phosphorylated AURKA were detected at a contractile ring/midbody during the first polar body extrusion. CPEB, localized in oocyte cytoplasm, was hyperphosphorylated during prophase/metaphase-I transition. Most CPEB degraded in metaphase-II oocytes and remnants remained localized in a contractile ring. Roscovitine, U0126, and metformin inhibited meiotic divisions; they all induced a decrease of CCNB1 and phospho-MAPK3/1 levels and prevented CPEB degradation. However, only metformin depleted AURKA. The Aurora kinase inhibitor VX680 at 100 nmol/L did not inhibit meiosis but led to multinuclear oocytes due to the failure of the polar body extrusion. Thus, in bovine oocyte meiosis, massive destruction of CPEB accompanies metaphase-I/II transition, and Aurora kinases participate in regulating segregation of the chromosomes, maintenance of metaphase-II, and formation of the first polar body.
aurora kinases, bovine, CPEB, gamete biology, kinases, meiosis, oocyte, oocyte development
In mammal ovaries, oocytes are blocked in prophase of the first meiotic division and become transcriptionally silent when they reach their full size and enter the preovulatory meiotic maturation process. Nuclear meiotic maturation of oocytes includes condensation of chromosomes, germinal vesicle breakdown (GVBD), progression through metaphase-I (MI), and an arrest at metaphase-II (MII). In cows, full-grown oocytes extracted from follicles larger than 2 mm are capable of resuming meiosis in vitro, and then of being fertilized and developing to the blastocyst stage [1]. The oocyte maturation process involves the coordinated action of several kinases. Mitogen-activated protein kinases (MAPKs) and maturation-promoting factor (MPF), a complex of cyclin B1 (CCNB1), and cell-cycle controller p34 kinase CDC2 (also known as CDK1, cyclin-dependent kinase 1) constitute the main signaling pathways of oocyte maturation. Two forms of serine-threonine MAPKs, MAPK1 and MAPK3 (also known as p42MAPK/ERK2 and p44MAPK/ERK1, respectively), are present in mammalian oocytes. They are activated, at least partially, by the upstream kinase MOS. In bovine oocytes, MPF and MAPK3/1 activation precedes GVBD during in vitro maturation (IVM) [2]. A significant increase in overall protein synthesis occurs in bovine oocytes during GVBD at 6–10 hours after the start of IVM; this synthesis then declines to reach basal level at the MII stage [3]. This process is concomitant with the polyadenylation of mRNA until the MII stage [4]. A number of proteins are synthesized de novo in bovine oocytes during IVM [5], including CCNB1 [6] and MOS [7]. In Xenopus and mouse oocytes, cytoplasmic polyadenylation and sequential translation of such mRNA are mediated by CPEB, a protein that binds to the cytoplasm polyadenylation element (CPE)—a U-rich cis-element present in the 3'-untranslated region of a number of mRNAs, such as MOS and CCNB1 [8, 9]. In Xenopus, activation of CPEB is triggered by the serine-threonine kinase AURKA [10,11], which is the homologue of yeast IpL1 and Drosophila Aurora, regulating the mitotic progression in eukaryotic cells [12]. However, it is not clear whether CPEB activation by AURKA triggers the further cascade of MAPKs phosphorylations leading to MPF activation and nuclear maturation [13], or whether AURKA itself needs the active MPF to be phosphorylated and activated, while these events are independent of the MOS/MAPK pathway [14]. In cytoplasm of mouse oocytes, AURKA is colocalized with CPEB, CPSF (cleavage and polyadenylation specificity factor), PAP (polyA polymerase), and maskin—factors known to control polyadenylation and translation—and was reported to mediate CPEB phosphorylation at the MI stage [15]. Phosphorylated CPEB activation stimulates both polyadenylation and translation of several mRNAs essential for oocyte maturation, including MOS, CCNB1, and SCP1/SCP3 (synaptonemal complex proteins 1 and 3) in mice [16]. AURKA may also be involved in microtubule assembly and nuclear activity in mouse oocytes [17].
In higher mammals, including domestic animals and humans, little is known about the involvement of Aurora kinases in meiosis. Nevertheless, in somatic cells, Aurora kinases have been identified as important regulators of mitotic divisions. Three Aurora kinases—AURKA, AURKB, and AURKC—exist in mammals. They share a similar kinase catalytic domain and their expression is cell cycle-regulated with a maximal protein level at the G2/M stage. Intracellular localization of Aurora kinases differs between AURKA and AURKB/AURKC [18]. AURKA is required for centrosome maturation and separation, bipolar spindle assembly, chromosomal alignment on the metaphase plate, and cytokinesis [21]; it localizes around centrosomes and spindle poles [19, 20]. AURKB is a chromosome passenger protein essential for chromosome condensation and cytokinesis completion; it localizes in chromosome kinetochores during prophase and metaphase and in the midbody during anaphase and telophase [22, 23]. AURKC, like AURKB, is a chromosome passenger protein [24]; it associates with AURKB and survivin [19, 25].
In cancer cells, Aurora proteins accumulate in cytoplasm [26, 27]. Aurora kinase activity is regulated by phosphorylation/dephosphorylation of several residues [28, 29]. Glycogen synthase kinase-3 (GSK-3) remains the only kinase identified as an upstream activator of Aurora [30]. Aurora kinases can also be activated by autophosphorylation when associated with some of their substrates, such as a motor-binding protein TPX2 (microtubule-associated protein homologue) for AURKA [31] or INCENP (inner centromere protein) for AURKB [29]. In turn, Aurora kinases phosphorylate a large number of proteins, including kinesin-like motor proteins, spindle apparatus proteins, kinetochore proteins, histones, and tumor suppressor proteins such as p53 (for review [18]).
We previously found that AURKA mRNA was highly expressed in bovine oocytes [32]. We also previously showed that AURKA protein level increased during IVM in bovine oocytes and that this increase was not affected by inhibition of MPF activity [33]. In pig oocytes, the Thr-phosphorylated AURKB level correlated with histone H3 phosphorylation on Ser10 [34]. In humans, AURKC mRNA has been reported to be overexpressed in oocytes as compared to cumulus cells [35].
The aim of this study was to establish the spatio-temporal expression patterns of three Aurora kinases in bovine oocytes during meiotic maturation and to clarify their roles in meiosis. First, we analyzed the relative expression of AURKA, AURKB, and AURKC mRNA by real-time RT-PCR and the location of the corresponding proteins by immunocytochemistry in bovine oocytes and preimplantation embryos. Second, mRNA and protein expression patterns of several main actors of the meiotic progression, such as CPEB, CCNB1, MOS, and CDC2, were studied in parallel with AURKA Thr-phosphorylation throughout IVM. Third, the effects of meiotic inhibitors roscovitine, U0126, and metformin on the levels of AURKA, phospho-MAPK3/1, CPEB, CCNB1, and CDC2 were assessed. Finally, the effects of VX680, the specific inhibitor of Aurora kinase activity, on meiosis progression were analyzed.
All procedures were approved by the Agricultural and Scientific Research Government Committees in accordance with the guidelines for Care and Use of Agricultural Animals in Agricultural Research and Teaching (approval A37801).
TCM199 culture medium, gentamicin, epidermial growth factor (EGF), fetal calf serum, MEK inhibitor U0126, and metformin were purchased from Sigma (Saint Quentin Fallavier, France). Cyclin-dependent kinase inhibitor roscovitine was provided by Dr. L. Meijer (Centre National de la Recherche Scientifique, Station Biologique de Roscoff, France). VX680 inhibitor of Aurora kinase activity was obtained from CAVA Technology (San Diego, CA).
Collection of Oocytes, Embryos, and Other Tissues
Cattle ovaries were collected from a slaughterhouse. Cumulus-oocyte complexes (COCs) were aspirated from 3- to 6-mm antral follicles. On average, 14 ± 4 good quality COCs were obtained from one slaughtered cow. COCs with more than three layers of compact cumulus cells surrounding the oocyte were selected and washed several times in TCM199/Hepes medium supplemented with 50 mg/L of gentamycin. Groups of 30–60 COCs were subjected to IVM in 500 µl of TCM199 medium supplemented with EGF (10 ng/ml) and 10% fetal calf serum at 39°C during 24 h in a humidified atmosphere containing 5% CO2.
In experiments using meiotic inhibitors, these substances were added to the culture medium just before IVM. U0126 and roscovitine were used at a final concentration of 100 µmol/L and 50 µmol/L, respectively. Metformin was added at 10 mmol/L, because it has been previously reported that this concentration does not affect cell viability [36]. We used VX680 at a concentration of 100 nmol/L—which is close to its half-maximal inhibitory concentration (IC50 15–113 nmol/L) measured on the proliferation of a variety of human cell types [37]—and also at 1 µmol/L, a concentration used in the past in studies on mouse and human cell lines (1 and 10 µmol/L) [38].
Oocytes were mechanically separated from cumulus cells either just after collections from ovaries (immature oocytes at germinal vesicle stage) or after 3, 6, 10, 14, and 22–24 h of IVM culture. Oocytes were rinsed in PBS, frozen in liquid nitrogen, and stored until RNA or protein extraction. The nuclear status of 20–50 oocytes was analyzed in each experimental situation. The meiotic status of oocytes at different times of maturation was established by lamin A/C immunofluorescence and chromatin labeling with Hoechst 33342 (Sigma, 1 µg/ml), followed by microscopic observation.
For in vitro preimplantation embryo production, COCs were subjected to IVM for 22 h, and subsequently to in vitro fertilization (IVF) and development, as previouosly described [32]. Groups of 10 or 20 embryos at one-cell, two-cell, four-cell, five- to eight-cell, morula, and blastocyst stages were collected and sorted, then either fixed for immunofluorescence analysis, or frozen in liquid nitrogen. Biopsies (0.5 g) from ovaries and testes were collected at the local INRA slaughterhouse from two adult cows and two bulls. Primary cultures of calf skin fibroblasts and cumulus cells were also used for protein extraction. All samples were frozen in liquid nitrogen and kept at –80°C before experiments.
RNA and cDNA Preparation and Analysis
Total RNA preparation. Total RNA was extracted from bovine oocytes, embryos, and biopsies of adult ovaries and testes using TriZol reagent as advised by the manufacturer (Invitrogen, Cergy Pontoise, France). To avoid contamination with genomic DNA, total RNA preparations were treated with RQ1 DNase (Promega) as described in the manufacturer's protocol.
AURKA, AURKB, and AURKC cDNA sequence analysis. Full-length Aurora A (gene AURKA) cDNA was obtained by using a SMART RACE cDNA Amplification Kit (Ozyme); 5' and 3' cDNA fragments were cloned into pCRII-dual promoter vector using the TA cloning kit (Invitrogen) and were sequenced by the Macrogen Company (Seoul, South Korea). Deduced cDNA and protein sequences were depositeded in Genbank (accession number DQ334808). Sequences for bovine Aurora B (AURKB) and Aurora C (AURKC) were found in GenBank (accession numbers NM_183084 and XM_870932, respectively). The sequences were analyzed using the software package from Infobiogen [39]. Alignments were performed using BLAST [40] and Multalin [41]. Deduced protein sequences were analyzed through the Interpro web site [42] and Simple Modular Architecture Research Tool (SMART) [43].
Reverse Transcription-PCR. Reverse transcription (RT) was performed on RNA amounts corresponding to 5 or 10 oocytes or embryos, or on 1 µg of total RNA from biopsies. Complementary DNA was extended from oligo(dT)15 primers during incubation for 1 h at 37°C by mouse Moloney leukaemia virus reverse transcriptase (Invitrogen), as described in the user's manual. For RT-PCR analysis, we used cDNA equivalent to 5% of one oocyte/embryo reversed RNA (1% of total RT reaction) as a template. For ovary and testis, 1/20 of the RT products were used (equivalent of 50 ng reversed RNA). PCR was performed using reagents from Interchim (Montluçon, France). In the negative control reaction, RNA amounts equivalent to one oocyte or 125 ng of tissue were directly subjected to PCR with specific primers. As positive control of cDNA quality, a β-actin-specific PCR was performed on all samples. Primer sequences are listed in Table 1. To confirm the specificity of amplified fragments, PCR products were cloned and sequenced.
|
Real-time RT-PCR. For real time RT-PCR analysis, cDNA was extended from the RNA of 10 oocytes by MMLV reverse transcriptase (Invitrogen) with 20 ng of oligo (dT)15 and 2 ng of 18S-antisense-RT primer per reaction. Real-time PCR was performed on MyiQ apparatus (Bio-Rad Laboratories, Marnes La Coquette, France) using a cDNA quantity equivalent to 5% of one oocyte/embryo per reaction. Reactions were performed in triplicate using a real-time PCR kit provided with SYBR Green fluorophore (Bio-Rad) according to the manufacturer's instructions and using a robotic distributor (Eppendorf). One picogram of luciferase mRNA was added to each group of 10 oocytes or embryos before RNA extraction and used as an external control of RNA extraction and cDNA quality. Four independent pools of RNA were analyzed for each stage of oocyte maturation and embryo development. For each considered gene, a standard curve was included, consisting of corresponding plasmid DNA fragments from 1 pg to 0.1 fg, purified with QIAquick PCR Purification kit (Qiagen). Correlation coefficients and PCR efficiencies were not less than 0.998 and 85%, respectively. Median values of the reaction triplicates were considered. Different approaches were employed for relative quantification of target polyA mRNA: 1) the relative abundance of AURKA, AURKB, and AURKC mRNA in oocyte, blastocyst, and testis was compared to the quantities of correspondent amplified products obtained from the same cDNA sample; 2) the relative abundances of target mRNA in oocytes at different maturation stages were evaluated in relation to that of 18S rRNA, considered as the internal reference gene since its level does not change during IVM in bovine oocytes [44]; 3) the relative abundance of AURKA mRNA in embryos at different development stages was calculated in relation to the external reference, luciferase polyA RNA, which was added before RNA extraction.
In all experiments, the mean value obtained for T0 (in immature oocytes) was considered as 1. Data were subjected to one-way ANOVA at a minimum level of significance of P < 0.05.
Antibodies. A monoclonal mouse antibody against recombinant full-length human Aurora A kinase has been produced and characterized previously [45]. Human Aurora B (ARK-2, H-75), CPEB (H-300), and MAPK42 (MAPK1 or ERK2, C14) rabbit polyclonal antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Phospho-Thr288 Aurora A rabbit polyclonal antibody was from Abcam (Cambridge, United Kingdom). Phospho-p44/42 MAPK rabbit polyclonal antibody (MAPK3/1 phosphorylated at Thr202/Tyr204) was purchased from Cell Signaling Technology (Danvers, MA). Alpha-tubulin (TUBA) monoclonal antibody (clone DM1) and Texas Red-conjugated goat antimouse antibody were from Sigma (Saint Quentin Fallavier, France). Antibodies against human Aurora C and phosho-Ser10 histone H3 were kindly provided by Dr. Y. Arlot (UMR6061 CNRS, France). Cyclin B1 monoclonal antibody (clone Ab-3), horseradish peroxidase (HRP)-conjugated rabbit antimouse, goat antirabbit, and goat antimouse IgG antibodies were purchased from Lab Vision Corporation (Fremont, CA). Lamin A/C monoclonal antibody was from Ozyme, (Saint Quentin Yvelines, France); Alexa Fluor488 goat antirabbit IgG and Alexa Fluor594 goat antimouse IgG were from Molecular Probes.
Western immunoblotting.
Groups of a definite number of bovine oocytes or embryos were frozen in 20 µl of Tris-saline-EGTA buffer (pH 7.5) supplemented with 2 mmol/L sodium ortovanadate and 1 µl/ml of protease inhibitor cocktail (Sigma), then thawed-frozen three times by rapid incubation in liquid nitrogen followed by immersion in a warm water bath at 30°C. In some experiments, the dephosphorylation of proteins on serine, threonine, and tyrosine residues was performed by using Lambda Protein Phosphatase (
-PPase, Sigma). Oocytes were placed in 20 µl of single-strength -PPase reaction buffer (50 mM Tris-HCl, 0.1 mM sodium-EGTA, 5 mM dithiothreitol (DTT), 0.01% Brij, 2 mM MnCl2), frozen-thawed three times, and then incubated with 400 U of -PPase for 1 h at 30°C. Before loading, concentrated reducing Laemmli buffer containing 80 mM DTT at final concentration was added to all protein extracts, and samples were boiled for 8 min.
Protein extracts were resolved on 10%–12% SDS-PAGE gels and transferred on nitrocellulose membranes. Blots were blocked with 5% dry milk in Tris buffered saline (TBS)/0.1% Tween 20 for 1 h at room temperature and probed with the various antibodies overnight at 4°C. Dilutions were 1/1000 for AURKA, phospho-MAPK3/1, total MAPK3/1, and TUBA antibodies; 1/200 for CDC2, CCNB1, and AURKB; 1/2000 for CPEB. After extensive washes in single-strength TBS/0.1% Tween, immunoreactivity was detected using the appropriate HRP-conjugated secondary antibodies (diluted 1/5000, incubated 1 h at room temperature) and revealed by enhanced chemiluminescence ECL Plus kit (Amersham Biosciences, Orsay, France) according to the manufacturer's instructions. Densitometry was performed by scanning the original radiographs and then analyzing the bands with Scion Image software Beta 4.0.2 (Fuji PhotoFilm). At least three blots were analyzed for each experimental condition.
Immunohistochemistry. Ovarian biopsies were fixed for 12 h in a solution containing 50% saturated picric acid, 3.7% formaldehyde, and 5% acetic acid. After serial dehydration steps, the samples were embedded in paraffin and serially sectioned at a thickness of 7 µm. Sections were deparaffined, rehydrated, microwaved for 5 min in antigen-unmasking solution (Vector Laboratories, Inc., AbCys, Paris, France), then left to cool at room temperature. After washing in a PBS bath for 5 min, sections were immersed in a peroxidase-blocking reagent for 10 min at room temperature to quench endogenous peroxidase activity (Dako Cytomation; Dako, Ely, United Kingdom). After three 5-min washes in a PBS bath, sections were blocked with 5% goat serum in PBS for 20 min. then incubated overnight at 4°C with PBS/0.1% BSA containing AURKA antibody (dilution 1/100). After three 10-min washes, sections were incubated for 30 min at room temperature with a biotinylated goat antimouse antibody. After serial washes, sections were stained for 10 min in a streptavidin peroxidase solution at room temperature as described in the kit manual (Lab Vision Corporation). Immunoreactivity was revealed by incubation at room temperature with 3,3'-diaminobenzidine (Lab Vision Corporation). The slides were counterstained with hematoxylin, then dehydrated and mounted in Depex (Sigma). Negative controls were performed by replacing primary antibodies with mouse IgG (Sigma) diluted in PBS/0.1% BSA at a final concentration of 2 µg/ml. Slides were observed using the Axioplan Zeiss transmission microscope.
Immunofluorescence of bovine oocytes and embryos. For fluorescent analysis, oocytes and blastocysts were fixed for 10 min in a solution containing 50% saturated picric acid, 3.7% methanol-stabilized formaldehyde (Interchim), and 5% acetic acid; samples were then subjected to four 15-min washes in PBS supplemented with 0.2% BSA and one 30-min wash in PBS/0.2% BSA/0.1% Triton. For AURKA detection during early embryogenesis, oocytes and embryos at one-cell, two-cell, 4-cell, 5- to 8-cell, morula, and blastocyst stages were fixed in cold methanol at –20°C for 2 h, rehydrated through serial ethanol baths, and washed three times in PBS with 0.2% BSA and 0.01% Tween20. Blocking was done in PBS/0.5% BSA supplemented with 10% goat inactivated serum for 2 h. Samples were incubated overnight with primary antibodies (diluted 1/100 for AURKA, AURKB, and AURKC; 1/200 to 1/500 for phospho-Thr AURKA; 1/200 for phospho-Ser10 histone H3 and Lamin A/C) at 4°C with constant shaking. At least four 30-min washes in PBS/0.2% BSA were done, and oocytes and embryos were incubated with corresponding secondary fluorochrome-conjugated antibody (diluted 1/200) for 1–2 h at room temperature. Five 20-min washes were then performed. Oocytes and embryos were put on slides and mounted with Mowiol solution supplemented with DABCO anti-fading agent and 1 µg/µl of Hoechst 33258 or DAPI (all from Sigma). Immunofluorescence was observed using an Axioplan Zeiss fluorescent microscope with appropriate filters.
Three Aurora Kinases Are Expressed in Bovine Oocytes
Using specific primers for Bos taurus Aurora kinase AURKA, AURKB, and AURKC genes, we detected by RT-PCR the cDNA fragments of expected sizes 234, 453, and 156 base pairs, respectively, in testis (Te), ovary (Ov), and full-grown immature oocytes (Oo) isolated from 3- to 6-mm antral follicles (Fig. 1A). AURKA was overexpressed in oocyte compared to AURKB and AURKC relative to β-actin amplification. To confirm these results, we quantified transcripts of the Aurora kinases by real-time PCR. AURKA mRNA in oocytes was about 20 times more abundant than AURKB and AURKC, while in testis its level was only about 2-fold higher (Fig. 1B). AURKA and AURKC mRNA levels decreased dramatically in blastocysts, while AURKB retained a similar level in oocytes and blastocysts (Fig. 1B).
|
Analysis of amino acid sequences showed that bovine AURKA, AURKB, and AURKC proteins shared a very similar C-terminal domain containing serine-threonine kinase catalytic domain and RXXL destruction-box (D-box), but differed in their noncatalytic N-terminal domain (see supplementary data, Figure 1A at www.biolreprod.org). Amino acid sequences of bovine Aurora kinases were highly homologous to those of human and mouse as shown for AURKA (see supplementary data, Figure 1B at www.biolreprod.org). Predicted molecular sizes of bovine Aurora proteins were 46 kDa, 40 kDa, and 35 kDa for AURKA, AURKB, and AURKC, respectively. To analyze their protein expression in bovine oocytes while taking into account the extreme similarity of bovine and human Aurora kinase amino acid sequences, we used antibodies to human Aurora proteins. Monoclonal antibody to human AURKA revealed a single protein at approximately 46 kDa in an extract prepared from 10 immature or mature oocytes (IO and MO) and in testis, but not in cumulus (CC) or fibroblast cells (Fb) (Fig. 1C). Human AURKB antibodies detected a protein of about 41 kDa in a protein extract prepared from fibroblast cells or from 50 mature oocytes but not in immature oocytes. In the same conditions, the antibody to human AURKC revealed the 35-kDa protein only in testis.
Thus, three Aurora kinases were expressed in oocytes, and Aurora A was the most abundant form in immature oocytes (both mRNA and proteins).
Localization of AURKA, AURKB, and AURKC in Bovine Oocytes During Meiosis and in Expanded Blastocysts
AURKA, Thr-phosphorylated AURKA, AURKB, and AURKC were visualized by indirect immunofluorescent detection in immature bovine oocytes at GV, MI, and MII oocytes and in embryos at the blastocyst stage (Fig. 2). In GV oocytes, AURKA was uniformly present in the whole cytoplasm (Fig. 2a). A subpopulation of AURKA phosphorylated on threonine residues was also detected, localized in numerous patches (Fig. 2b). No significant staining was observed when rabbit or/and mouse nonspecific IgG was used instead of primary antibodies (Fig. 2d). In MI and MII oocytes, AURKA (Fig. 2, f and k) and phospho-Thr -AURKA (Fig. 2, g and l) were dispersed through the cytoplasm. Sometimes AURKA was concentrated around the MII plate (Fig. 2k, MII inset); phospho-Thr-AURKA immunoreactivity was not detected at that location (Fig. 2l, MII inset). Phospho-Thr-AURKA was concentrated at a contractile ring in the area that separates from the polar body (PB) in oocytes at telophase-1 and MII stages (Fig. 2m; see also Fig. 5E, c). AURKA was not detected in the spindle poles in MI-MII oocytes (Fig. 2, f and g, k–m). However, in the expanded blastocyst, AURKA was preferentially concentrated in the spindle poles in metaphase, anaphase, and in the newly forming poles in prophase cells (Fig. 2p, insets M, An, and P, respectively). Thr-phosphorylated AURKA was clearly detected in the spindle poles during metaphase in the blastocyst (M, asterisk-labeled inset in Fig. 2p).
|
|
AURKB was practically undetectable in the cytoplasm of immature oocytes (Fig. 2 c). In MI and MII oocytes, AURKB was detected on the chromosomes (Fig. 2, h and n) like in mitotic blastocyst cells (Fig. 2q, insets M). AURKB was also detectable in the cytoplasm of MII and to a lesser extent in MI oocytes. However, AURKB was neither detected in germinal vesicle nor in the polar body chromatin (Fig. 2, c and n). In contrast, AURKB was clearly associated with the contractile ring separating the polar body in MII oocyte (Fig. 2n, inset PB) and with the contractile ring/midzone between blastomers during telophase/cytokinesis (Fig. 2q, inset MZ).
AURKC also showed a cytoplasmic localization throughout meiosis (Fig. 2, e, i, and o). In MII oocytes, the protein was concentrated between the polar body and what might be the contractile ring (Fig. 2o, inset PB). No significant AURKC specific labeling was detected in blastocyst cells.
Taken together, all three known Aurora kinase genes were expressed in bovine oocytes during progression of meiosis. AURKA protein showed the most intriguing expression pattern because it is quite different from that in the somatic cells. In contrast, AURKB in oocytes was associated with the metaphase chromosomes and with the midbody, in accordance with the established pattern in mitotic cells. The faint level of Aurora C expression did not allow its study at the protein level. Thus, we concentrated on the analysis of Aurora A.
AURKA Expression Pattern in the Ovary During Folliculogenesis, in Fertilized Oocytes, and in Early Embryos
We analyzed the localization of AURKA in the oocytes throughout folliculogenesis on the paraffin-embedded sections of bovine ovaries (Fig. 3A). All main steps of folliculogenesis were found in the sections of the adult cow ovary: 1) primary follicles in which the oocyte is surrounded by a single layer of cells; 2) secondary and tertiary follicles where the oocyte has more than one granulosa layers; and 3) small antral (0.3–2 mm), antral (2–6 mm), and preovulatory follicles (>6 mm in diameter) in which the oocyte is surrounded by cumulus cells to form a COC. AURKA was poorly detectable in primary Fig. 3A, a), secondary (Fig. 3A, b), and tertiary preantral follicles (Fig. 3A, c). In small antral follicles of about 0.3–0.6 mm in diameter, AURKA immunostaining clearly appeared in the cytoplasm of oocytes that measured more than 80% of the size of a full-grown oocyte (Fig. 3A, d). AURKA was also detected in some cumulus, granulosa, and theca cells in actively growing follicles (Fig. 3A, d and f) and to a lesser extent in large follicles (Fig. 3A, g). No staining was detected without the first antibody (Fig. 3A, e). Along with follicular antrum growth, AURKA staining increased in the oocyte, except in the nucleus (Fig. 3A, f and g). In all oocytes collected from antral follicles, AURKA was detected exclusively in the cytoplasm.
|
Western blots showed that AURKA protein level increased during IVM, then remained stable until the morula stage, and was decreased significantly in the expanded blastocyst (Fig. 3B, black bars). In parallel, we quantified AURKA polyadenylated mRNA in oocytes and embryos (Fig. 3B, gray bars). The level of polyA mRNA encoding AURKA was relatively stable in immature, mature, and fertilized oocytes; it decreased slightly and progressively during the three first cleavages (two-cell, four-cell, and five- to eight-cell), and fell tremendously in embryos at the morula and blastocyst stages.
Immunofluorescence showed that AURKA was equally abundant in the cytoplasm of unfertilized and fertilized oocytes and in teo-cell and four-cell embryos (Fig. 3C), and was not significantly concentrated at the spindle poles. We observed the concentration of AURKA in the spindle poles only from the third embryo cleavage (five- to eight-cell). At this stage and onward, AURKA appeared concentrated at the spindle poles and at the nearest spindle microtubules in all mitotic blastomers going through pro-metaphase, metaphase, and anaphase (eight-cell and morula in Fig. 3C; blastocyst in Fig. 2p).
Thus, AURKA accumulated at a very high level in oocyte cytoplasm during final follicular growth and maturation. AURKA was Thr-phosphorylated in oocytes. AURKA protein was not degraded until activation of the embryonic genome. AURKA was definitely relocated to the spindle poles from the eight-cell stage on (i.e., from the moment of the maternal-embryo transition). To determine the role of AURKA in oocytes, we followed the expression of this kinase simultaneously with several main actors implicated in the meiotic progression in vitro.
Expression of AURKA, CPEB, MOS, CCNB1, and CDC2 in Bovine Oocytes During IVM
Once collected from the antral follicles and put into IVM culture medium, bovine oocytes passed through the sequential stages of nuclear maturation, finally attained MII, and were stopped at that stage. Under these conditions, the recovered oocytes were in meiotic prophase and retained the intact GV for at least 6 h, as confirmed by lamin A/C immunostaining (not shown). The first GVBD occurred after approximately 6 h of IVM, and the MI stage was achieved at 10 h of IVM. At 22 h, most of the oocytes were at MII (Fig. 4A). We quantified polyA mRNA levels of AURKA, CCNB1, MOS, CDC2, and CPEB in oocytes at 3, 6, 10, and 22 h of IVM by using real-time PCR performed on oligo(dT)-extended cDNA (Fig. 4B). In this experiment, the reference immature oocytes (Fig. 4B; 0 h, white bars) were retrieved from ovaries transported at 4°C and were manipulated at 4°C up to RNA extraction to prevent cytoplasmic polyadenylation of oocyte mRNA before IVM. Total RNA level did not significantly change during IVM as verified by 18S rRNA quantification. Levels of polyA AURKA, CCNB1, CDC2, and MOS mRNA increased during the first 3 h of IVM and, from this time onward, were relatively stable up to 10 h of IVM. After 22 h of IVM, the AURKA polyA mRNA level did not change as compared to 0 h, while the CDC2 level was significantly diminished. CCNB1, CPEB, and MOS mRNA levels were slightly decreased in MII oocytes (Fig. 4B).
|
We quantified AURKA, CCNB1, CPEB, and CDC2 protein levels in the oocytes at the same time of IVM (Fig. 4C). AURKA and CCNB1 increased significantly during maturation. Two closely migrated shifted AURKA bands were detected in the oocytes throughout IVM (Fig. 4D), which might have been a consequence of AURKA phosphorylation. In this experiment, special sampling and SDS-PAGE procedures were used, as described previously [46]. In contrast with AURKA, CCNB1 was not detectable before IVM, and its appearance coincided with the phosphorylation of MAPK3/1 at 3 h of IVM (Fig. 4C). CDC2 protein was maintained at a relatively constant level throughout IVM (Fig. 4C). CPEB level was stable in immature oocytes up to GVBD (6 h of IMV), then declined at 10 h and was mostly degraded at 22 h of IVM. The slower migrating CPEB-immunoreactive shifted band was detected at 6 h and 10 h of IVM (Fig. 5A). We performed the -PPase treatment of the oocytes at 10 h of IVM and demonstrated that this slower migrating band corresponded to the hyperphosphorylated form of CPEB (Fig. 5B). Double immunofluorescence confirmed that both AURKA and CPEB localized in the oocyte cytoplasm (Fig. 5C). CPEB was not detected as intensely in the cytoplasm of mature oocytes (MII) as compared to immature oocytes (GV). However, in MII oocytes, CPEB was clearly detected in the region of the polar body separation, which might be a contractile ring/midbody (Fig. 5C, inset PB).
Effect of Roscovitine, U0126, and Metformin on Oocyte Maturation
Addition of roscovitine (50 µmol/L), a specific inhibitor of CDC2/MPF activity, stopped the oocytes at the GV stage after 22 h of IVM, but neither significantly influenced the accumulation of AURKA (Fig. 5D) nor AURKA phosphorylation, as detected by immunofluorescence (Fig. 5E, d) and gel shift (supplementary data, Fig. 2). The phosphorylation of MAP kinases was inhibited by the addition of the specific MAPK kinase inhibitor U0126 (100 µmol/L). The U0126-treated oocytes proceeded beyond GVBD but did not progress to MII after 22 h of IVM (Fig. 5D). Nevertheless, accumulation of AURKA was not significantly affected, and we did not observe any loss of phospho-Thr- AURKA labeling in U0126 oocytes (Fig. 5E, f). In contrast, the addition of metformin (10 mmol/L), an insulin-sensitizing agent and an activator of AMPK (adenosine monophosphate-activated kinase), stopped the oocytes progression before GVBD and significantly decreased AURKA synthesis and the level of phospho-AURKA immunofluorescence (Fig. 5D and 5E, e). The level of phospho-MAPK3/1 was significantly reduced in the presence of roscovitine, U0126, or metformin. In contrast, CPEB level was not diminished after 22 h of IVM in the oocytes treated with roscovitine, U0126, or metformin, although in control IVM experiments, without inhibitors or with DMSO, CPEB was largely degraded in mature oocytes (Fig. 5D).
Thus, in bovine oocytes, inhibition of MPF or MAPK activation during IVM stopped meiosis before the first division. These treatments had no effect on AURKA accumulation and activation, but AURKA depletion by metformin coincided with a stop in progresion at the GV stage. Interestingly, all tested inhibitors decreased the level of phospho-MAPK3/1 at 22 h of IVM, and CPEB was neither hyperphosphorylated nor degraded.
Effect of Aurora Kinase Inhibitor VX680 on Oocyte Maturation
We performed IVM either in the presence of VX680, a small molecule inhibiting Aurora kinase activity, or with DMSO as a control. VX680 is a potent inhibitor of AURKA, AURKB, and AURKC with apparent inhibition constant values of 0.6, 18.0, and 4.6 nmol/L, respectively [37]. After IVM, oocytes were denuded from the cumulus cells and their nuclear status was monitored by the chromatin Hoechst staining. When 1 µM concentration of VX680 was used, 72.4% of oocytes were stopped before the MI stage, whereas 85% of control oocytes were already in MII. In VX680-treated oocytes, no phospho-Ser10 histone H3 staining was detected in a compact clump of condensed chromosomes, whereas in control oocytes the phospho-Ser10 histone H3 was associated with the metaphase chromosomes and with the polar body chromatin (Fig. 6A). Practically all the VX680-treated oocytes passed through GVBD, while no lamin A/C staining was detected around the chromatin (not shown). In the presence of a lower concentration of VX680 (100 nmol/L), all oocytes continued meiosis from MI onward and phospho-Ser10 Histone H3 was detected associated with the chromatin (Fig. 6A). However, many oocytes showed an abnormal meiosis and supernumerary chromatin structures after treatment with the lower concentration of VX680 (Table 2, Fig. 6B). Immature bovine COCs were cultured in IVM medium supplemented or not with 100 nM VX680 for 14 or 24 h. After 14 h of IVM, most of the control oocytes were at the MI stage (Fig. 6B, a), while in VX680-treated oocytes, several chromatin groups were detected in the cytoplasm (Fig. 6B, b–d). Abnormally large, non-extruded polar bodies (Fig. 6B, b and c) or supernumerary chromatin structures (Fig. 6B, b) were observed in oocytes treated with 100 nM VX680. The chromosomes were sometimes completely disorganized and dispersed throughout the cytoplasm (Fig. 6B, d). After 24 h of IVM, most of the control oocytes were at the MII stage (Fig. 6B, e). In contrast, multiple chromatin structures, similar to activated pronuclei or to non-extruded double polar bodies, were detected in oocytes treated with 100 nM VX680 (Fig. 6B, f–h). Quantification of several polyA transcripts revealed a decrease after 3 h of IVM of only MOS mRNA level in oocytes treated with VX680 as compared to control oocytes (Fig. 6C). By Western blot, no differences were detected in AURKA, CDC2, or phospho-MAPK3/MAP1 protein levels between groups treated with 100 nM VX680 and control groups (Fig. 6D). In contrast, while CPEB was detectable in control oocytes after 14 h of IVM, CPEB was already degraded in oocytes treated with 100 nM VX680.
|
|
Aurora Kinases in Bovine Oocytes: AURKB and AURKC Were Localized as in Mitotic Cells While AURKA Had a Different Expression Pattern
We have shown experimentally that three Aurora kinases are expressed in bovine oocytes, and that AURKA is more prevalent in immature oocytes. AURKA and AURKB protein levels increased during IVM. In human oocytes, Assou et al. [35] have shown that the level of AURKC mRNA is higher in MI and MII compared to immature oocytes. We found that in mature oocytes AURKB, AURKC, and the active Thr-phosphorylated AURKA were all concentrated in the contractile ring/midbody zone, localized at the furrow separating the first polar body. In fact, the contractile ring is an actomyosin-based structure essential for the completion of cytokinesis; it is formed on the central spindle in early anaphase with the bundling of overlapping microtubules. These bundles become compacted and mature into the midbody (for review [47]). AURKB is a component of the central spindle assembly complex, and it phosphorylates another component, kinesin-6 protein MKLP1, in vitro and in vivo [48]. Recent studies have reported the colocalization of AURKC with AURKB/survivin complex in HeLa cells during mitosis [25]. Thus, the detection of AURKB and AURKC in the midzone between the ooplasm and the first polar body is consistent with their expression pattern in the mitotic cells. More surprisingly, phospho-AURKA was also found in this zone, and this indicates that AURKA may have downstream substrates at this site. Thus, in bovine oocytes, Aurora kinases may participate in regulating the first polar body extrusion.
AURKB in bovine oocytes was also shown to be bound to the metaphase chromosomes in meiosis I and II, but not to the chromatin in the germinal vesicle or in the polar body. This suggests that the role of AURKB in the correct alignment and segregation of the chromosomes during meiotic divisions is similar to its role in mitotic cells. These results are in concordance with data showing that in pig oocytes AURKB is not involved in chromosome condensation during first meiosis while the chromosomes are condensed despite the inhibition of AURKB and histone H3 phosphorylation [34].
AURKA protein was detected in oocytes in correlation with the appearance of the antrum and final follicular growth. The signals for antrum formation are not well understood, but in vitro studies performed with rodent follicles have shown that FSH, LH, KL (kit ligand), and EGF are possible candidate regulators (for review [49]). In cows, antral follicles up to 5 mm in size are not yet dependent on gonadotropins. Progesterone and androgens are likely the main steroid hormones produced by theca cells, since granulosa cells lack the expression of P450 aromatase, converting steroids to estrogens [49]. In frog oocytes, progesterone induces AURKA accumulation and activation [11, 46]. It is not known whether progesterone or other factors induced the initial translation of AURKA in bovine oocytes. However, our data suggest that accumulation of AURKA in oocyte cytoplasm starts a long time before ovulation, and the activated form of this kinase was detected in GV oocytes retrieved from antral follicles. In Xenopus oocytes, the upper band from two closely migrating bands revealed by the AURKA antibody corresponds to Thr-phosphorylated AURKA [30, 50] and was detected at GVBD [50, 51]. In cows, the shifted AURKA band was already detectable before GVBD, and this was confirmed by immunofluorescent detection of Thr-phosphorylated AURKA in immature oocytes. Phosphorylated AURKA was visible on the surface and dispersed in the ooplasm of immature oocytes but was absent from the spindle zone in MII oocytes. However, the pan-AURKA eventually concentrated around the MII-aligned chromosomes. Perhaps inactive AURKA participates in preventing premature mitotic divisions of unfertilized oocytes and thus maintains MII arrest. The absence of active maternal centrosomes in the spindle poles in bovine and human oocytes could also explain why AURKA is not concentrated at the spindle poles [52]. In contrast, AURKA was concentrated at the spindle poles in mitotic cells in bovine embryos from the eight-cell stage (i.e., at the stages when the embryo genome is already active [53]), and AURKA was Thr-phoshorylated there. AURKA protein level was stable up to the morula stage and diminished significantly in the blastocyst. The same has been observed in Xenopus embryos, where the protein level is stable during several cleavages while AURKA activity oscillates. In contrast, in somatic cells, AURKA protein level peaks at the G2/M stage and is degraded during the exit from mitosis. Indeed, destruction of AURKA and several other mitotic regulators, such as cyclins, requires an activator of the ubiquitin ligase APC/C (anaphase-promoting complex/cyclosome) called Cdh1 (for review [54]). However, in Xenopus eggs and early embryos, Cdh1 is missing, and consequently AURKA is not degraded during the initial cleavages [55]. APC/C recognition sequences, the destruction-boxes, the D-box, and the A-box are all conserved in Xenopus and human Aurora genes [56, 55]. A putative A-box required for Cdh1-dependent destruction was found within the bovine AURKA amino acid sequence, similar to the human, mouse, and frog gene homologues (Supplementary data, Fig. 1B). In mice, Cdh1 is already present in MII oocytes, and some APCCdh1 substrates begin to degrade from the second polar body extrusion [57]. Taking into account that embryos are transcriptionally active at the late one-cell stage in mice, at the eight-cell stage for bovine embryos, and after 12 cleavages at the midblastula stage for frogs, we hypothesize that in bovine oocytes, maternal AURKA must be produced and activated in order to accomplish its role up to maternal-embryo transition, after which it might be degraded, probably via the APCCdh1 mechanism.
Therefore, a high concentration of active Aurora A in the cytoplasm of fully grown immature bovine oocytes is probably required for the progression of meiosis, similar to the situation in frogs and mice (for review [18]).
AURKA Synthesis and Phosphorylation Were Independent from MAPK and MPF Activation and Were Not Required for CPEB Hyperphosphorylation and Degradation During IVM
An increase in MOS, CCNB1, and CDC2 polyadenylated mRNA levels was observed during the first 3 h of IVM and the subsequent appearance of CCNB1 protein. Our results are in agreement with reported RACE-PAT analysis in bovine oocytes, which shows that the CCNB1 mRNA polyA tail is progressively elongated during the first 10 h of IVM and that CCNB1 is synthesized from the transcripts with a long polyA [58]. MOS protein is produced in bovine oocytes from around the onset of meiotic resumption, and becomes phosphorylated at GVBD [59]. In pig oocytes, although MOS transcripts are polyadenylated, their translation is repressed before GVBD and then increases steadily from MI to MII [60]. According to our data, AURKA protein was already present in the active phosphorylated form and was localized together with CPEB in the cytoplasm of immature oocytes (i.e., before the massive translation of CCNB1 and MOS). AURKA level increased up to MI, and CPEB became hyperphosphorylated at that time. MOS, CCNB1, CDC2, and AURKA mRNAs contain the CPE in their 3'-UTRs. Thus, AURKA may be implicated in the regulation of polyadenylation-dependent translation of these transcripts during oocyte maturation via its participation in the activation of CPEB as in Xenopus and mouse [11, 13, 15, 30, 61]. CPEB is a highly conserved RNA-binding protein [62] that is a general regulator of meiosis. In CPEB knock-out mice, oogenesis and spermatogenesis are disrupted at pachytene [63]. In Xenopus laevis, phosphorylation of CPEB by Aurora A was reported to trigger the interaction of CPEB and CPSF, which in turn binds to the AAUAAA sequence and recruits poly(A) polymerase to the end of mRNA and therefore regulate the polyadenylation-dependent translation of CPE-containing transcripts (for review [8]). It has also been reported that the overexpression of recombinant AURKA in human and rat cell lines induces the phosphorylation of CPEB and promotes the polyadenylation of CCNB1 and CDK1 (CDC2) mRNAs [64]. In mouse oocytes, AURKA has been reported to phosphorylate CPEB first at the leptotene stage of prophase and then at MI, initially triggering the translation of SCP1/SCP3 and then MOS and CCNB1, respectively [61]. We found that AURKA was active in prophase oocytes and was thus potentially capable of activating CPEB. Moreover, hyperphosphorylated CPEB was observed during the GVBD-MI transition, before the degradation of CPEB and the relocation of remnants to the contractile ring/midzone in MII oocytes. The fact that CPEB colocalizes with phospho-AURKA at the contractile ring/midzone may indicate the local translation/activation of MOS, CCNB1/CDC2 or other substrates. In synchronized HeLa cells, AURKA has been shown to participate in the recruitment of the CCNB1/CDC2 complex to centrosomes, where MPF becomes activated and commits cells to mitosis [65]. In Xenopus, once the polyadenylation takes place during oocyte maturation, most CPEB is destroyed; all that remains stable is confined to the animal pole blastomeres in embryos, where it is strongly associated with the spindle and centrosomes and is involved in the localization of CCNB1 mRNA to the mitotic apparatus [66]. Intriguingly, in bovine oocytes, all tested meiotic inhibitors blocked CPEB hyperphosphorylation and degradation, although the levels of AURKA and Thr-phosphorylation were not significantly affected, as observed during treatment with roscovitine or U0126. One of the factors responsible for CPEB degradation in frog eggs is active CDC2 [67]. Indeed, when we blocked CDC2 activity by roscovitine, CPEB was not degraded. The arrest before MI and the decrease of MAPK3/1 phosphorylation were the most common effects of roscovitine, U0126, and metformin on oocytes after 22 h of IVM. Therefore, active MAPK may also be involved in CPEB hyperphosphorylation and its consequent degradation. In fact, a recent study reported that MAPK activation is required for the phosphorylation of CPEB during meiosis in Xenopus oocytes, and a lower level of MAPK activation was detected before MOS synthesis [50].
A significant decrease in AURKA level was observed in metformin-treated oocytes with subsequent meiotic arrest at the GV stage, and this was accompanied by the complete depletion of CCNB1 synthesis and a significant decrease in MAPK3/1 phosphorylation. Metformin inhibits protein synthesis through the eEF2 kinase/eEF2 axis and/or the p70S6 kinase pathway [68]. In bovine and pig oocytes, metformin also provokes meiotic arrest [69, 70]. Interestingly, metformin has recently been reported to be an AMPK-dependent growth inhibitor for breast cancer cells [71], in which AURKA is also known to be overexpressed [26, 72]. AMPK is present in bovine oocytes and cumulus cells [36, 69], and it would be interesting to know whether there is a functional relationship between AMPK and AURKA. Although nonspecific, the depletion of AURKA by metformin in bovine oocytes resulted in the arrest of meiotic progression. In Xenopus oocytes, the ablation of AURKA protein by the microinjection of a specific antibody also blocks meiosis, but the oocytes are arrested at MI [73].
In contrast to CCNB1 and MOS, CDC2 was already present in immature bovine oocytes, and no variation in protein quantity was observed during maturation, consistent with previous reports [6, 74]. This suggests that CDC2 mRNA should be polyadenylated and translated during oogenesis earlier than CCNB1 and MOS. In human cells, with DNA damage and arrested at G2/M, the overexpression of AURKA triggers an override of the G2/M arrest through a CDK1/CDC2 reactivation, suggesting the existence of a retroactive control loop between AURKA and CDK1 [75]. In frog oocytes, active CDC2 has been shown to be necessary for AURKA activation [14, 46, 76]. Our results showed that AURKA was phosphorylated before the activation of MPF and MAPK3/1, and before GVBD. Moreover, the inhibitors for CDC2/MPF or MAPK3/1 activity (roscovitine or U0126, respectively) did not change the AURKA and phospho-Thr-AURKA expression pattern during IVM. These data confirmed that AURKA accumulation and phosphorylation were independent of MPF or MAPK activation in bovine oocytes, as has been shown in other models [30].
Effect of the Aurora Kinase Activity Inhibitor VX680 on the Meiotic Progression of Oocytes
VX680 blocks the kinase activity of all three Aurora kinases, although it shows the greatest selectivity for AURKA [37]. In human cells, VX680 inhibits cell proliferation, induces DNA endoreduplication and tetraploidization, and leads to apoptosis [37, 77]. In our experiments, oocytes treated with 1 µM VX680 resumed GVBD but did not proceed to MI. Since the histone H3 was not phosphorylated in oocytes treated with 1 µM VX680, AURKB was probably inhibited at that concentration, together with AURKA and AURKC. Similarly, in pig oocytes, Aurora kinase inhibitor ZM447439, more specific for AURKB, prevents both the activation of AURKB and phosphorylation of histone H3 on Ser10. ZM447439-treated oocytes are arrested just after GVBD in the late diakinesis stage [34]. Although AURKB is not required for chromosome condensation [34], it is required for maintenance of chromatin condensation in Xenopus [78] and surf clam oocytes [79], and thus AURKB activity might be critical for the correct MI/MII transition.
At a concentration of 100 nmol/L, VX680 neither significantly affected the Ser10-phosphorylation of histone H3 nor blocked progression through meiosis. Thus, AURKB may not be inactivated. However, abnormal meiotic events—including chromosome misalignment, chromatin decondensation, and formation of pronucleus-like structures—were observed, and these might be the consequences of AURKA inactivation. Oocytes treated with 100 nM VX680 seemed to accelerate meiosis since polar bodies (often nonextruded) were already detected, whereas most of the control oocytes were only at MI after 14 h of IVM. In fact, pronucleus-like chromatin structures were detected within the cytoplasm of oocytes treated with 100 nM VX680 whereas the control oocytes were at MII. The same observation has been made in Xenopus oocytes in which AURKA activity is inhibited [80]. While the MOS polyA mRNA level was diminished in oocytes treated with 100 nM VX680, it could be presumed that due to the lower level of MOS protein, the oocytes might enter parthenogenesis. Indeed, oocytes derived from MOS-deficient knock-out mice failed to arrest at MII and underwent parthenogenetic activation [81]. Similarly, bovine and pig oocytes, in which the endogenous MOS mRNA was depleted with double-strand or antisense RNA microinjections, were parthenogenetically activated [82–84]. VX680 also disturbed cytokinesis and therefore treatment resulted in oocyte polyploidy. While CPEB might already be active before retirement from the follicles, the polyadenylation of CCNB1 could be initiated before IVM, as reported in bovine species [58], and therefore maturation progressed and GVBD occurred. However, MOS protein must be synthesised after GVBD in order to maintain MII arrest as a component of the cytostatic factor [85, 86]. Perhaps while kinase activity of AURKA was at least partially inhibited by VX680, MOS might not be synthesized at a high enough level at this time, and therefore oocytes exited from MII and began DNA replication.
In conclusion, three Aurora kinases and CPEB are expressed during meiosis in bovine oocytes. AURKA, AURKB, and AURKC show distinct expression patterns during IVM but localize together in the midzone during the first polar body extrusion. The inhibitor of Aurora kinase activity affected correct meiotic division and a polar body extrusion. Therefore, Aurora kinases may play a role in chromosome segregation as in mitotic cells, but also an oocyte-specific role in MII maintenance. AURKA is phosphorylated in immature oocyte cytoplasm and in the region of the polar body midzone where CPEB also localizes. Nevertheless, the presence of active AURKA was not sufficient to produce the hyperphosphorylation and the destruction of CPEB during MI/MII transition. Altogether, these observations indicate the putative involvement of AURKA in regulating polyadenylation-dependant translation in bovine oocytes, although this role will have to be demonstrated by further studies.
ACKNOWLEDGMENTS
We thank Barbara Schmaltz, Juan Traverso, Gael Ramé, and Abdulrahman Aldarwich for technical assistance and Philippe Monget for helpful discussions. We are grateful to Alice Fatet for careful correction of the English text.
FOOTNOTES
1Supported by grants of ANR and APIS-GENE as part of the OVOGENAE program. C.P. was supported by the Cancéropôle Grand Ouest, the LNCC, and the ARC. 
Correspondence: 2Svetlana Uzbekova, UMR 6175 INRA-CNRS-Université de Tours Station Physiologie de la Reproduction et des Comportements, 37380 Nouzilly, France. FAX: 33 2 47 42 77 43; e-mail: uzbekova{at}tours.inra.fr
Received: 20 February 2007.
First decision: 17 March 2007.
Accepted: 27 July 2007.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  K. Dieterich, R. Zouari, R. Harbuz, F. Vialard, D. Martinez, H. Bellayou, N. Prisant, A. Zoghmar, M. R. Guichaoua, I. Koscinski, et al. The Aurora Kinase C c.144delC mutation causes meiosis I arrest in men and is frequent in the North African population Hum. Mol. Genet., April 1, 2009; 18(7): 1301 - 1309. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Dupont, B. D. Bavister, D. R. Armant, and C. A. Brenner Rhesus macaque embryos derived from MI oocytes maturing after retrieval display high rates of chromosomal anomalies Hum. Reprod., April 1, 2009; 24(4): 929 - 935. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
-tubulin (TUBA) was performed as a control of protein loading.
