ABSTRACT
Oocytes of almost all vertebrates become arrested at metaphase II to await fertilization. Arrest is achieved with the participation of a protein complex known as cytostatic factor (CSF) that stabilizes histone H1 kinase activity. MOS and mitogen-activated protein kinase (MAPK) are important components of CSF. Strain LT/Sv mice, and strains related to LT/Sv, produce a high percentage of atypical oocytes that are arrested at metaphase I when normal oocytes have progressed to metaphase II. The potential role of MOS in metaphase I arrest was investigated using strain LT/Sv and LT-related recombinant inbred strains, LTXBO and CX8-4. MOS and MAPK are produced and functional in maturing LT oocytes. Two experimental paradigms were used to reduce or delete MOS in LT oocytes and assess effects on metaphase I arrest. First, sense and antisense Mos oligonucleotides were microinjected into metaphase I-arrested oocytes. Antisense, but not sense, Mos oligonucleotides promoted the activation of metaphase I-arrested oocytes. Second, mice carrying a Mos null mutation were crossed with LT mice, the null mutation was backcrossed three times to LT mice, and Mos+’∼ N3 mice were intercrossed to produce Mos−/−, Mos+/− and Mos+/+ N3F1 mice. Oocytes of all three Mos genotypes of N3F1 mice sustained meiotic arrest for 17 hours indicating that metaphase I arrest is not initiated by a MOS-dependent mechanism. However, unlike Mos+/+ and Mos+/− CX8-4 N3F1 oocytes, metaphase I arrest of Mos−/− CX8-4 N3F1 oocytes was not sustained after 17 hours and became reversed gradually. These results, like the antisense Mos oligonucleotide microinjection experiments, suggest that MOS participates in sustaining metaphase I arrest in LT oocytes.
INTRODUCTION
In mammalian females, prophase of meiosis I usually begins in the fetal period but progression to cell division is blocked at the diplotene stage; oocytes resume meiosis after they acquire and activate molecules that drive completion of the G2 to metaphase transition. Meiotic resumption is first manifested by disassembly of the nuclear membrane (germinal vesicle breakdown, GVB) and condensation of chromosomes. Normally, metaphase I is transient in fully grown oocytes and followed by anaphase I and telophase, segregation of one set of homologous chromosomes into the first polar body, and progression of meiosis through metaphase II without an intervening interphase. Meiosis is arrested again at metaphase II until penetration of a spermatozoan activates the oocyte and triggers entry into anaphase II and the completion of meiosis II. However, the progression of meiosis in many fully grown oocytes of strain LT/Sv mice is abnormal and arrested precociously at metaphase I (Eppig and Wigglesworth, 1994; Eppig et al., 1996; Kaufman and Howlett, 1986; West et al., 1993). Moreover, many metaphase I-arrested LT/Sv oocytes undergo spontaneous parthenogenetic activation and embryo development (Eppig et al., 1996). When this activation and development occurs within the ovaries, parthenogenetic embryos can develop into ovarian teratomas (Stevens and Varnum, 1974). Analysis of the mechanism of metaphase I arrest in LT oocytes is therefore essential for understanding the etiology of ovarian teratocarcinogenesis in this strain and may also reveal new information on the regulation of the metaphase I to anaphase I transition in normal oocytes.
The eukaryotic cell cycle is regulated by oscillations in the activity of M-phase Promoting Factor (MPF) consisting of a catalytic subunit, p34cdc2 and a regulatory subunit, cyclin B. Entry into metaphase is driven by the activation of MPF, and anaphase is correlated with cyclin B destruction and decreased MPF activity; see Jacobs (1992), Murray (1995), and Norbury and Nurse (1992) for reviews. Oscillations in MPF activity also drive meiosis in mammalian oocytes (Choi et al., 1991; Fulka et al., 1992; Hampl and Eppig, 1994, 1995; Hashimoto and Kishimoto, 1988). The characteristic decrease in MPF activity seen in normal oocytes coincident with anaphase I does not occur in metaphase I-arrested LT/Sv oocytes. Instead, activity continues to increase. Likewise, the amount of cyclin B present in metaphase I-arrested oocytes continues to increase, in contrast to its degradation in normal oocytes progressing to metaphase II (Hampl and Eppig, 1995). These results suggested that metaphase I arrest in LT/Sv oocytes is the result of continued MPF activity, sustained, at least in part, by restricted degradation of cyclin B.
MPF activity is necessary to maintain metaphase II arrest in oocytes, and the function of a multi-component complex, known as cytostatic factor (CSF), is required to sustain MPF activity. CSF activity is the coordinated function of at least two proteins, MOS and mitogen-activated protein kinase (MAPK). MOS, the product of the Mos proto-oncogene, is transcribed as a dormant mRNA throughout mouse oocyte growth (Gebauer et al., 1994; Goldman et al., 1987; Keshet et al., 1988; Mutter and Wolgemuth, 1987). Translation of Mos mRNA occurs in mouse oocytes mainly after the resumption of meiosis and MOS accumulates during oocyte maturation (Paules et al., 1989). MOS is not necessary for meiotic resumption in mouse oocytes, as it is in frog oocytes (Sagata et al., 1989; Yew et al., 1992), since GVB occurs in oocytes of Mos−/− mice produced by homologous recombination (Colledge et al., 1994; Hashimoto et al., 1994; Verlhac et al., 1996). In contrast, MOS is required for metaphase II arrest in mouse oocytes since this does not occur in most oocytes from Mos−/− mice (Araki et al., 1996; Colledge et al., 1994; Hashimoto et al., 1994; Verlhac et al., 1996). MOS is also essential for MAPK activity since there is little or no MAPK activity detectable in oocytes of Mos−/− mice (Araki et al., 1996; Choi et al., 1996; Verlhac et al., 1996). Thus, MOS probably functions to maintain metaphase II arrest by promoting MAPK activity, which in turn may either inactivate the cyclin B degradation system, prevent an increased rate of degradation, or both. The critical consequence of either mechanism is sustained high MPF activity and maintenance of metaphase II arrest.
Given this role of MOS in metaphase II arrest, the hypothesis underlying this study is that MOS participates in initiating or sustaining metaphase I arrest in LT/Sv oocytes. To test this idea, the potential role of MOS in metaphase I arrest was investigated using oocytes of strain LT/Sv mice and a recombinant inbred strain, LTXBO, derived from strains LT/Sv and C57BL/6J. Since equivalent results were obtained using both LT/Sv and LTXBO oocytes, both strains are referred to generically as LT. MOS and MAPK are both produced and functional in maturing LT oocytes. Two experimental paradigms were used to reduce or delete MOS in LT oocytes and assess the effects on metaphase I arrest. First, sense (control) and antisense Mos oligonucleotides were microinjected into metaphase I-arrested oocytes. Antisense, but not sense, Mos oligonucleotides promoted the activation of metaphase I-arrested oocytes. Second, a Mos null mutation was crossed to LT mice and then backcrossed three times to LT mice; Mos+/+ N3 mice were intercrossed to produce Mos−/−, Mos+/− and Mos+!+ N3F1 mice. An identical backcross experiment was performed using another LT-related strain, CX8-4, in which many fully grown oocytes become arrested at metaphase I, but the frequency of spontaneous parthenogenetic activation is very low (Eppig et al., 1996). This is a useful model to study the possible participation of MOS in the maintenance of metaphase I arrest, since progression of oocytes beyond metaphase I in the absence of MOS cannot be due to LT-related spontaneous parthenogenetic activation. Surprisingly, oocytes from all of these backcross mice exhibited a very high occurrence of metaphase I arrest, implying that the initiation of metaphase I arrest is not initially MOS dependent. However, metaphase I arrest of Mos−/− N3F1 oocytes was temporary and was gradually reversed. This observation, in combination with the reversal of metaphase I arrest by antisense Mos oligonucleotides, suggests that MOS participates in sustaining metaphase I arrest in LT oocytes.
MATERIALS AND METHODS
Mice
Mice were raised in our research colony at The Jackson Laboratory. For most experiments, strain LT/Sv mice were used, but when specified, strain LTXBO oocytes were used in their place. LTXBO is a recombinant inbred strain derived from LT/Sv and C57BL/6J by L. C. Stevens and D. S. Varnum at The Jackson Laboratory (unpublished). Like one of its progenitors, LT/Sv, LTXBO oocytes exhibit a high frequency of metaphase I arrest, and the frequency of spontaneous parthenogenetic activation and ovarian teratocarcinogenesis is higher than in LT/Sv, possibly because a gene(s) inherited from C57BL/6J promotes the development of parthenote embryos (Eppig et al., 1996). Since equivalent results were obtained here using either LT/Sv or LTXBO oocytes, the term LT will be used hereafter in reference to either strain. A Mos null mutation was introduced into the LTXBO mice, as well as into another recombinant inbred strain CX8-4 derived from BALB/c and C58/J (the progenitor strains of LT/Sv) (Eppig et al., 1996). CX8-4 oocytes show the same frequency of metaphase I arrest as LT/Sv oocytes, but a low frequency of parthenogenetic activation (Eppig et al., 1996). (C57BL/6J × SJL/J)FI oocytes (hereafter referred to as F1 oocytes) were used as normal control oocytes.
Collection and maturation of oocytes in vitro
Mice were injected with 5 IU equine chorionic gonadotropin (eCG; Diosynth, Oss, Holland) at 20–22 days of age, and fully grown germinal vesicle (GV)-stage oocytes were isolated 44-48 hours after eCG injection as described previously (Eppig et al., 1996). Unless specified, the cumulus cells around oocytes were removed immediately after isolation from the follicles by drawing the complexes in and out of small bore pipettes. Oocyte maturation medium was Minimum Essential Medium (MEM) prepared with Earle’s balanced salt solution, essential amino acids (GIBCO, Grand island, NY), 0.23 mM pyruvic acid, 75 mg/l penicillin-G, 50 mg/l streptomycin sulfate, 10 gM disodium ethylenediaminetetraacetic acid (Sigma Chemical Co. St. Louis, MO) and 3 mg/ml bovine serum albumin (crystallized, ICN Immunochemicals, Lisle, IL). Cultures were placed in a modular incubation chamber (Billups-Rothenberg, Del Mar, CA), equilibrated to 5% CO2, 5% O2 and 90% N2, and incubated at 37°C. At the end of culture, the oocytes were examined with a Wild M5A stereomicroscope to determine whether they had produced a polar body, an indication of progression to metaphase II. The oocytes were then either frozen for biochemical analyses or mounted on slides, fixed with 4% formaldehyde and stained with lacmoid before examination using a phase-contrast optics.
Immunoblotting
100 oocytes were used for each LT oocyte sample. However, because LT oocytes are slightly larger than F1 oocytes (Eppig and Wig-glesworth, 1994), the total volume of oocytes was equalized between samples by using 16% more F1 oocytes than LT oocytes (Chesnel and Eppig, 1995). Oocytes were washed twice with ice-cold phosphate-buffered saline (PBS) containing 3 mg/ml polyvinylpyrrolidone (Sigma), 2 μg/ml aprotinin, 5 μg/ml leupeptin, 70 μg/ml L-1-chloro-3-[4-tosylamido]-4-phenyl-2-butanone, 0.7 μg/ml pepstatin and 1 mM 4-(2-aminoethyl)-benzenesulfonylfluoride (Boehringer Mannheim, Indianapolis, IN). Oocytes were then collected in microcentrifuge tubes and stored at −70°C until use.
Oocytes were lysed in 2× Laemmli sample buffer (Laemmli, 1970) and heated to 100°C for 3 minutes. Samples were run on 10% SDS–PAGE and proteins were transferred to nitrocellulose membrane (Hybond-ECL; Amersham, Arlington Heights, IL) using a semidry blotting apparatus (OMNI BIO, Czech Republic). Membrane-bound proteins were probed with rabbit polyclonal IgG raised against mouse MOS subdomain III, residues 148-177 (UBI, Lake Placid, NY). Membranes were treated with horseradish-peroxidase-labelled antirabbit immunoglobulins (GIBCO) and the activity was visualized using the ECL western blotting detection system (Amersham). Each experiment was repeated at least three times and a representative is shown.
MAPK activity assay
Oocytes were lysed in the Laemmli sample buffer as described above, except that 60 and 70 oocytes were used for LT mice and FI mice, respectively. Samples were subjected to SDS–PAGE in a 10% acrylamide gel containing 0.4 mg/ml myelin basic protein (MBP; UBI), a preferred substrate of MAPK. In-gel kinase assay was performed as described previously (Gotoh et al., 1991; Kameshita and Fujisawa, 1989); following denaturation and renaturation processes, the gel was incubated with 2.5 μ Ci/ml [γ-32P]ATP and radioactivity incorporated was visualized using an image analyzer (Fuji Bio-Imaging Analysis System; Fuji Medical Systems USA, Stanford, CT) or X-ray film (Dupont NEN). Each experiment was repeated at least three times and a representative is shown.
Oocyte microinjection
Oligonucleotide phosphorothioates purified through a successive HPLC and gel filtration were purchased from Genosys Biotechnologies, Inc. (The Woodlands, TX) and were resuspended in Ca2+/Mg2+-free PBS to achieve a final concentration of 1 mg/ml. The 15-base sequences of the targeted region of endogenous Mos mRNA and its antisense complement were as follows (5’-3’): sense, CTCCACTCA-CAAAGC; antisense, GCTTTGTGAGTGGAG (Gebauer et al., 1994; O’Keefe et al., 1989). About 10 pl of oligonucleotide solution was injected into the cytoplasm of metaphase I-arrested LT/Sv oocytes matured for 14-15 hours. Microinjection was performed in a microdrop of MEM (pH 7.2) prepared with 20 mM Hepes, 5 mM NaHCO3 and 20% fetal bovine serum. After microinjection, oocytes were washed in maturation medium and cultured for 8–10 hours. At the end of culture, oocytes were mounted on slides, processed and examined as described above.
Production of Mos null LT mice
The recombinant inbred strain LTXBO, derived from LT/Sv and C57BL/6J, was used to receive the Mos null mutation. A Mos null male (Colledge et al., 1994) was mated with LTXBO females, and resulting heterozygotes were used for backcross to LTXBO mice. N3 backcross mice were mated to obtain Mos null mice. Mos+/+, Mos+/−, and Mos−/− mice were identified by polymerase chain reaction (PCR) using tail DNA. The primers were (5′-3′: neo 784 (GGGTGGA-GAGGCTATTCGGCTAT); neo 323 (GAAGAACTCGTCAAGAAG-GCGATAGAA); Mos P5 (GCCCTAGTGGTATGTTTTCTCCAC); and Mos P6 (GGAAGTGCCTCAAGTATTCCCTAG). These generated a 0.75 kb neo and a 0.6 kb Mos PCR product in heterozygotes, but only the neo product in Mos−/− mice. Mos+/+, Mos+/− and Mos−/− N3F1 mice were also produced from strain CX8-4 in the same manner.
Oocytes were collected and cultured as described above except that cumulus cells were not removed during maturation. After 17 or 24 hours of maturation, the oocytes were denuded of cumulus cells, mounted on slides, processed and examined as described above.
Statistical analysis
Maturation and activation data are presented as mean percentages of at least three independent experimental replicates in which all groups were included in each experiment; variation between experiments is illustrated using the standard error of the mean. For evaluation of the differences between the groups, data were subjected to arcsin transformation and ANOVA. When a significant F ratio was identified, groups were compared using Fisher’s protected least significant difference posthoc test using Statview for Macintosh (Abacus Concepts, Inc., Berkeley, CA); when P< 0.05, the difference was considered significant.
RESULTS
General characteristics of nuclear maturation of oocytes from FI and LT mice
As shown previously (Eppig and Wigglesworth, 1994), cumulus cell-denuded FI oocytes showed typical maturation kinetics in vitro; GVB occurred about 1 hour after the beginning of culture and emission of a polar body by 12 hours (data not shown). In contrast, although LT oocytes also underwent GVB with normal kinetics, many oocytes failed to produce a polar body at the normal time and were arrested at metaphase I. Approximately 50% of the cumulus cell-denuded oocytes still had not produced a polar body by 18 hours of culture (Eppig and Wigglesworth, 1994).
MOS production and MAPK activation in LT oocytes
The MOS antibody recognized a band of about 38× 103Mrin the lysates of maturing F1 and LT oocytes after 12 hours of culture (Fig. 1A). It was not detectable in samples prepared from either GV-stage control oocytes or Mos−/− oocytes, suggesting that the recognized protein was MOS (Fig. 1A). The amount of MOS accumulated during maturation decreased dramatically after fertilization (Fig. 1B). Accumulation of MOS appeared equivalent throughout oocyte maturation in both control and LT/Sv oocytes (Fig. 1C). This accumulation occurred regardless of polar body emission in LT/Sv oocytes (not shown).
MOS in oocytes during meiotic maturation. (A) Lysates of 100 oocytes were subjected to SDS–PAGE followed by western blotting with MOS polyclonal antibody. Lane 1, GV-stage LT/Sv oocytes; lanes 2, 3 and 4, oocytes obtained from control (B6SJLF1), Mos−/− and LT/Sv mice, respectively, and matured for 12 hours in vitro. Mobilities of standard proteins of the indicated molecular masses are shown on the left. (B) The change in relative amount of MOS in control (B6SJLF1) oocytes after fertilization. Lysates of 100 oocytes were subjected to SDS–PAGE followed by immunoblotting using anti-MOS polyclonal antibody. Lane 1, GV-stage oocytes; lane 2, metaphase II oocytes matured in vitro for 17 hours; lanes 3 and 4, oocytes 4 and 6 hours, respectively, after insemination of metaphase II-arrested oocytes. The image of the blot was digitized and analyzed by ImageQuaNT software (Molecular Dynamics, Inc.). Bars below the blot represent the relative quantified volume of the bands. (C) The change in relative amounts of MOS during meiotic maturation in control (B6SJLF1) and LT/Sv oocytes. Lysates of oocytes were subjected to SDS-PAGE followed by immunoblotting using MOS polyclonal antibody. Lanes 1, 2 and 3, F1 oocytes matured for 6, 12 and 18 hours, respectively; Lanes 4, 5 and 6, LT/Sv oocytes matured for 6, 12 and 18 hours, respectively. By 12 hours, control (B6SJLF1) oocytes typically had extruded the first polar bodies, thus such oocytes were collected at 12 and 18 hours. However, only oocytes without polar bodies (metaphase I-arrested) were selected in LT/Sv group at all time points. For the immunoblot shown, 99 F1 oocytes and 85 LT/Sv oocytes were used per lane to equalize the total oocyte volume.
MOS in oocytes during meiotic maturation. (A) Lysates of 100 oocytes were subjected to SDS–PAGE followed by western blotting with MOS polyclonal antibody. Lane 1, GV-stage LT/Sv oocytes; lanes 2, 3 and 4, oocytes obtained from control (B6SJLF1), Mos−/− and LT/Sv mice, respectively, and matured for 12 hours in vitro. Mobilities of standard proteins of the indicated molecular masses are shown on the left. (B) The change in relative amount of MOS in control (B6SJLF1) oocytes after fertilization. Lysates of 100 oocytes were subjected to SDS–PAGE followed by immunoblotting using anti-MOS polyclonal antibody. Lane 1, GV-stage oocytes; lane 2, metaphase II oocytes matured in vitro for 17 hours; lanes 3 and 4, oocytes 4 and 6 hours, respectively, after insemination of metaphase II-arrested oocytes. The image of the blot was digitized and analyzed by ImageQuaNT software (Molecular Dynamics, Inc.). Bars below the blot represent the relative quantified volume of the bands. (C) The change in relative amounts of MOS during meiotic maturation in control (B6SJLF1) and LT/Sv oocytes. Lysates of oocytes were subjected to SDS-PAGE followed by immunoblotting using MOS polyclonal antibody. Lanes 1, 2 and 3, F1 oocytes matured for 6, 12 and 18 hours, respectively; Lanes 4, 5 and 6, LT/Sv oocytes matured for 6, 12 and 18 hours, respectively. By 12 hours, control (B6SJLF1) oocytes typically had extruded the first polar bodies, thus such oocytes were collected at 12 and 18 hours. However, only oocytes without polar bodies (metaphase I-arrested) were selected in LT/Sv group at all time points. For the immunoblot shown, 99 F1 oocytes and 85 LT/Sv oocytes were used per lane to equalize the total oocyte volume.
A well-documented role for MOS is to phosphorylate MAPK kinase, leading to activation of MAPK (Nebreda and Hunt, 1993; Posada et al., 1993; Shibuya and Ruderman, 1993). To determine if MOS is functional in LT/Sv oocytes, MAPK activity was assessed using an in-gel kinase assay method. As shown in Fig. 2, MAPK is clearly activated in LT oocytes after meiotic resumption. There was intense incorporation of radioactivity at the two regions of 44× 103 and 42× 103Mr, which correspond to ERK1 and ERK2, respectively (Verlhac et al., 1993). Thus, MOS is both present and functional in LT oocytes.
MAPK activity in control (F1) and LT oocytes detected using in-gel kinase assay. Lysates of 70 and 60 oocytes of FI and LT/Sv mice, respectively, were subjected to the electrophoresis on SDS-polyacrylamide gel containing MBP. Lanes 1 and 2, GV-stage FI and LT/Sv oocytes, respectively; lane 3, FI oocytes, with polar bodies, matured for 12 hours; lane 4, LT/Sv oocytes, without polar bodies, matured for 12 hours; lane 5, FI oocytes, with polar bodies, matured for 18 hours; lanes 6 and 7, LT/Sv oocytes, without and with polar bodies, respectively, matured for 18 hours. Mobilities of standard proteins of the indicated molecular masses are shown on the left. As expected, no MAPK activity was detected in GV-stage oocytes (0 hours). MAPK activity was equivalent in FI and LT oocytes.
MAPK activity in control (F1) and LT oocytes detected using in-gel kinase assay. Lysates of 70 and 60 oocytes of FI and LT/Sv mice, respectively, were subjected to the electrophoresis on SDS-polyacrylamide gel containing MBP. Lanes 1 and 2, GV-stage FI and LT/Sv oocytes, respectively; lane 3, FI oocytes, with polar bodies, matured for 12 hours; lane 4, LT/Sv oocytes, without polar bodies, matured for 12 hours; lane 5, FI oocytes, with polar bodies, matured for 18 hours; lanes 6 and 7, LT/Sv oocytes, without and with polar bodies, respectively, matured for 18 hours. Mobilities of standard proteins of the indicated molecular masses are shown on the left. As expected, no MAPK activity was detected in GV-stage oocytes (0 hours). MAPK activity was equivalent in FI and LT oocytes.
Microinjection of antisense Mos oligonucleotides into LT oocytes
Normally, spontaneous parthenogenetic activation of LT oocytes occurs only rarely after maturation of cumulus cell-denuded oocytes (Eppig, 1982). When cumulus cell-denuded LT oocytes were microinjected with antisense Mos oligonucleotides at metaphase I, 50% formed a pronucleus during 8–10 hours of culture (Fig 3). Injection of antisense Mos oligonucleotides often caused pronuclear formation immediately after meiosis I. Metaphase II-arrested mouse oocytes become activated after inhibition of protein synthesis (Fulka et al., 1994). It has also been suggested that antisense oligonucleotides are likely to cause at least partial degradation of many non-targeted RNAs in Xenopus oocytes (Woolf et al., 1992). Therefore, to test whether antisense oligonucleotides inhibit protein synthesis generally, metaphase I-arrested oocytes were radiolabeled with [35 S] methionine (1 mCi/ml) in Whitten’s medium (Whitten, 1971) for 2 hours beginning at 2 hours after microinjection. Incorporation of 35S into radiolabeled trichloroacetic acid-precipitable material per oocyte was determined for 4 replicates of 20 or 15 oocytes for each group. The mean values (cpm) ± s.e.m./oocyte were 1411.0±28.5 and 1270.9±69.2 for the uninjected oocytes and oocytes microinjected with antisense oligonucleotides, respectively. When paired /-test was applied, the difference was not significant (P=0.16). Thus, the effect of oligonucleotides appears not to be caused by non-specific inhibition of protein synthesis.
Effect of microinjection of antisense Mos oligonucleotides into LT/Sv oocytes. Microinjection was performed on metaphase I-arrested oocytes matured for 14-15 hours. Oocytes were fixed for examination 8-10 hours after injection. Bars indicate the percentage of oocytes of the stages shown on the bottom: MI, metaphase I; MII, metaphase II; PN, pronucleus. The data are presented as mean percentages of three independent experimental replicates; variation between experiments is illustrated using the standard error of the mean. The number of oocytes included in each group of experiments 1-3 is: uninjected control = 21, 14, 25; sense = 12, 10, 32; antisense = 14, 11,22. *A significant difference (P<0.05) from the sense oligonucleotide-injected group.
Effect of microinjection of antisense Mos oligonucleotides into LT/Sv oocytes. Microinjection was performed on metaphase I-arrested oocytes matured for 14-15 hours. Oocytes were fixed for examination 8-10 hours after injection. Bars indicate the percentage of oocytes of the stages shown on the bottom: MI, metaphase I; MII, metaphase II; PN, pronucleus. The data are presented as mean percentages of three independent experimental replicates; variation between experiments is illustrated using the standard error of the mean. The number of oocytes included in each group of experiments 1-3 is: uninjected control = 21, 14, 25; sense = 12, 10, 32; antisense = 14, 11,22. *A significant difference (P<0.05) from the sense oligonucleotide-injected group.
Mos null LTXBO and CX8-4 oocytes
The experiments described above using antisense oligonucleotides suggests that MOS participates in the maintenance of metaphase I arrest in LT oocytes. However, it was possible that oocyte activation could have resulted from various nonsequence-specific mechanisms caused by microinjection of oligonucleotides. Therefore, to further assess the relationship between MOS function and metaphase I arrest, a Mos null mutation, produced by homologous recombination (Colledge et al., 1994), was backcrossed to LTXBO mice to the N3 generation. N3Mos+/− heterozygotes were intercrossed to produce Mos+/+, Mos+/− and Mos−/− N3F1 mice. The Mos null mutation was also backcrossed to another recombinant inbred strain CX8-4. Oocytes from CX8-4 mice show the same frequency of metaphase I arrest as LT/Sv oocytes, but a very low frequency of parthenogenetic activation (Eppig et al., 1996). Thus, if metaphase I arrest is reversed in Mos null CX8-4 backcross oocytes, the reversal cannot be due to LT-related spontaneous parthenogenetic activation. As expected, MAPK was inactive in maturing oocytes of N3F1Mos−/− mice (Fig. 4). Since genes that promote metaphase I arrest are dominant (Eppig et al., 1996; West et al., 1993), oocytes from Mos+/− N3F1 mice became arrested at metaphase I (∼90%) (Fig. 5A,C).
Absence of MAPK activity in Mos null LTXBO backcross N3F1 oocytes. Lysates of 60 Mos+/− or Mos−/− oocytes per lane were subjected to the electrophoresis on SDS-polyacrylamide gel containing MBP. Lanes 1 and 2, GV-stage Mos+/− and Mos−/− oocytes, respectively; lanes 3 and 4, Mos+/− and Mos−/− oocytes, respectively, matured for 10 hours.
Absence of MAPK activity in Mos null LTXBO backcross N3F1 oocytes. Lysates of 60 Mos+/− or Mos−/− oocytes per lane were subjected to the electrophoresis on SDS-polyacrylamide gel containing MBP. Lanes 1 and 2, GV-stage Mos+/− and Mos−/− oocytes, respectively; lanes 3 and 4, Mos+/− and Mos−/− oocytes, respectively, matured for 10 hours.
Metaphase I arrest in cumulus cell-enclosed Mos null LTXBO backcross N3F1 oocytes (A,B) and Mos null CX8-4 backcross N3F1 oocytes (C,D). Cumulus cell-enclosed oocytes were matured for 17 hours (A,C) or 24 hours (B,D). Bars indicate the percentage of oocytes of the stages shown on the bottom. Data are presented as the mean percentage of four independent experimental replicates ± s.e.m. The number of oocytes included in each group of experiments 1–4 is: (A) Mos+/− = 62, 55, 67, 47; Mos−/−= 54, 44, 55, 33; (B)Mos+/− = 53, 50, 53, 44; Mos−/−= 65, 48, 61, 31; (C)Mos+/− = 38, 53, 48, 56; Mos−/− = 37, 47, 78, 38; (D) Mos+/− = 39, 48, 61, 57; Mos−/−= 35, 32, 82, 38. *The Mos−/−group is different (P<0.05) from the Mos+/− group of the same stage of nuclear maturation in that panel. Oocytes that were at the stages of germinal vesicle, anaphase I and telophase I were included in the percentagecalculation, but not shown in the figure.
Metaphase I arrest in cumulus cell-enclosed Mos null LTXBO backcross N3F1 oocytes (A,B) and Mos null CX8-4 backcross N3F1 oocytes (C,D). Cumulus cell-enclosed oocytes were matured for 17 hours (A,C) or 24 hours (B,D). Bars indicate the percentage of oocytes of the stages shown on the bottom. Data are presented as the mean percentage of four independent experimental replicates ± s.e.m. The number of oocytes included in each group of experiments 1–4 is: (A) Mos+/− = 62, 55, 67, 47; Mos−/−= 54, 44, 55, 33; (B)Mos+/− = 53, 50, 53, 44; Mos−/−= 65, 48, 61, 31; (C)Mos+/− = 38, 53, 48, 56; Mos−/− = 37, 47, 78, 38; (D) Mos+/− = 39, 48, 61, 57; Mos−/−= 35, 32, 82, 38. *The Mos−/−group is different (P<0.05) from the Mos+/− group of the same stage of nuclear maturation in that panel. Oocytes that were at the stages of germinal vesicle, anaphase I and telophase I were included in the percentagecalculation, but not shown in the figure.
Surprisingly, however, the same percentage of metaphase I arrest was observed in the Mos−/−N3F1 oocytes cultured for 17 hours (Fig. 5A,C). After 24 hours of culture, as typical of cumulus cell-enclosed LTXBO oocytes, 41% and 46% of Mos +/− and Mos−/− LTXBO backcross oocytes, respectively, formed a pronucleus indicative of parthenogenetic activation (Fig. 5B). Thus, abrogation of metaphase I-arrest and activation could not be attributed to the absence of MOS. In contrast, over 90% of Mos+/− CX8-4 backcross oocytes remained arrested at metaphase I after 24 hours of culture, while only 15% of Mos−/− oocytes were arrested and 64% of them formed a pronucleus (Fig. 5D). Results from Mos+!+ oocytes were essentially the same as Mos+/− oocytes (not shown). These observations clearly indicate that MOS function participates in the maintenance of metaphase I arrest.
DISCUSSION
A high percentage of LT oocytes become atypically arrested at metaphase I when most normal mouse oocytes have progressed to metaphase II (Eppig and Wigglesworth, 1994; Eppig et al., 1996; Kaufman and Howlett, 1986; O’Neill and Kaufman, 1987; West et al., 1993). MOS is a component of CSF, a protein complex essential for maintaining metaphase II arrest in vertebrate oocytes (Masui, 1991; Sagata, 1996, 1997). Evidence is presented here that MOS is produced by LT oocytes. Moreover, since normal levels of MAPK activity were found in maturing LT oocytes, MOS itself is probably not defective. Furthermore, evidence is presented that MOS does not instigate metaphase I arrest in LT oocytes, but participates in sustaining prolonged metaphase I arrest.
During maturation, the amount of MOS accumulated in LT oocytes appears similar to that of normal oocytes. MOS in LT oocytes is probably functional since MAPK appeared activated normally. Taken together, these results suggest that there are no abnormalities in the temporal production or function of MOS per se that contribute to the atypical maturation of LT oocytes.
To determine whether MOS participates in sustaining metaphase I arrest in LT oocytes as it does in sustaining metaphase II arrest in normal oocytes, antisense and sense (control) Mos deoxyoligonucleotides were microinjected into metaphase I-arrested LT oocytes. Antisense, but not sense, oligonucleotides promoted the progression of meiosis and activation of metaphase I-arrested LT oocytes. This finding suggested that MOS participates in the maintenance of metaphase I arrest. To test this hypothesis further, a Mos null mutation (Colledge et al., 1994) was backcrossed to LTXBO and CX8-4 mice and then intercrossed to produce N3F1Mos+/+, Mos+/− and Mos−/− mice. Since microinjection of antisense Mos oligonucleotides promoted activation, it was surprising that the Mos−/−N3F1 oocytes cultured for 17 hours remained arrested at metaphase I. Thus, MOS does not instigate metaphase I arrest in LT oocytes. Although MOS deficiency could cause some delay of polar body emission (Hashimoto et al., 1994), the frequency of metaphase I arrest of control Mos−/− oocytes (Colledge et al., 1994) was normal (15%) at 17 hours of maturation (data not shown). But the observation that Mos−/− CX84 backcross oocytes gradually became activated after 17 hours confirms the conclusion from the oligonucleotide microinjection experiment that MOS participates in the prolonged maintenance of metaphase I arrest. Thus, the mechanism of the primary block of cell cycle is separable from MOS function to sustain metaphase I arrest.
It is not known why LT oocytes initially become arrested at metaphase I. MPF activity remains high and cyclin B continues to accumulate in LT oocytes when MPF activity decreases and cyclin B is degraded in normal oocytes progressing from metaphase I (Hampl and Eppig, 1995). However, cyclin B degradation and reduced MPF activity are probably not sufficient for entry into anaphase I. Anaphase I differs from both mitotic anaphase and meiotic anaphase II; homologous chromosomes segregate during anaphase I and sister chromatids separate during mitotic anaphase and meiotic anaphase II. Entry into mitotic anaphase or meiotic anaphase II requires processes independent of decreased MPF activity before separation of sister chromatids is possible (Wells, 1996). In anaphase I, there may be mechanisms that initiate separation of homologs conjoined by chiasmata that are independent of decreased MPF activity and these may be defective in LT oocytes. For example, treatment of mouse oocytes with etoposide, an inhibitor of topoisomerase II, causes metaphase I arrest by preventing the transition to anaphase I (Fulka et al., 1994). In this case, a metaphase I checkpoint may detect a failure in topoisomerase-dependent resolution of chiasmata-linked homologous chromosomes and, therefore, not activate anaphase effector mechanisms. Alternatively, LT oocytes may have a defective metaphase I-specific checkpoint that fails to assess the establishment of a normal metaphase I and thus may not be able to signal separation of homologous chromosomes or the activation of the cyclin B degradation pathway.
Regardless of the mechanism instigating metaphase I arrest in LT oocytes, it is transient since homologous chromosomes do separate upon parthenogenetic activation (Eppig et al., 1977; Eppig and Eicher, 1983), which promotes completion of meiosis I, first polar body emission and pronuclear formation (Eppig et al., 1996). Based on the findings reported here, we hypothesize that there is a 2-step mechanism for metaphase I arrest in LT oocytes. The first step reflects the fundamental lesion in LT oocytes; the step that initiates metaphase I arrest. The second step is the maintenance of metaphase I arrest by CSF activity, which presumably functions when the transient mechanism instigating metaphase I arrest is abrogated. As a hypothetical example, metaphase I arrest in LT oocytes could be initiated by an aberrant topoisomerase II with a retarded catalytic activity that causes a delay in the initiation of a metaphase I checkpoint effector system. By the time that the metaphase I checkpoint can initiate anaphase I, CSF activity develops to the point where it prevents cyclin degradation, MPF activity remains high and metaphase I arrest is sustained.
The 2-step mechanism described above depicts intrinsic events occurring within LT oocytes that initiate and sustain metaphase I arrest. However, this mechanism can be influenced by extrinsic factors since cumulus cells increase the frequency of metaphase I arrest (Eppig and Wigglesworth, 1994). Cumulus cells may affect either the instigation or prolongation of metaphase I arrest, or both. Nevertheless, that metaphase I arrest cannot be sustained without MOS was clearly indicated by the high frequency of parthenogenetic activation observed in Mos null CX8-4 backcross oocytes. Moreover, spontaneous parthenogenetic activation of LT oocytes occurs only rarely after maturation of cumulus cell-denuded oocytes (Eppig, 1982) and only when oocytes are arrested in metaphase I (Eppig et al., 1996; Maleszewski and Yanagimachi, 1995). Both of these were true for LTXBO N3F1 oocytes in this study, regardless of whether they were Mos+/− or Mos−/−. Cumulus cells appear to destabilize metaphase I arrest in LTXBO oocytes; a destabilization that promotes activation. It is critically important to determine the mechanism that instigates the metaphase I arrest since this may be the primary lesion in LT oocytes necessary for subsequent parthenogenetic activation and teratoma development. Moreover, these aberrant meiotic processes in LT oocytes could provide an avenue for revealing the mechanisms that regulate the progression of normal meiosis in mammalian oocytes including the existence of a metaphase I checkpoint.
ACKNOWLEDGEMENTS
This research was supported by a grant (CA 62392) from the National Cancer Institute to JJE. Scientific Resources of The Jackson Laboratory are supported in part by a Cancer Center Core Grant (CA 34196) from the National Cancer Institute. We are grateful to Drs. Mary Ann Handel, Barbara Knowles and Bermseok Oh for their helpful comments during the preparation of this manuscript.