Meiotic Arrest in the Mouse Follicle Maintained by a Gs Protein in the Oocyte

See allHide authors and affiliations

Science  23 Aug 2002:
Vol. 297, Issue 5585, pp. 1343-1345
DOI: 10.1126/science.1073978


The mammalian ovarian follicle consists of a multilayered complex of somatic cells that surround the oocyte. A signal from the follicle cells keeps the oocyte cell cycle arrested at prophase of meiosis I until luteinizing hormone from the pituitary acts on the follicle cells to release the arrest, causing meiosis to continue. Here we show that meiotic arrest can be released in mice by microinjecting the oocyte within the follicle with an antibody that inhibits the stimulatory heterotrimeric GTP–binding protein Gs. This indicates that Gs activity in the oocyte is required to maintain meiotic arrest within the ovarian follicle and suggests that the follicle may keep the cell cycle arrested by activating Gs.

Oocytes within mammalian ovarian follicles begin meiosis during embryogenesis but then arrest at prophase of meiosis I until luteinizing hormone acts on the follicle to cause meiosis to resume (1). Maintaining this arrest in fully grown oocytes depends on the presence of the surrounding follicle (Fig. 1, A and C); removing the oocyte from the follicle reinitiates meiosis. However, it is unknown how the follicle cells communicate with the oocyte to keep the cell cycle arrested. Signaling depends on maintaining a high level of adenosine 3′,5′-monophosphate (cAMP) in the oocyte, but where the cAMP comes from and how the follicle cells regulate its level is unclear (1). One hypothesis is that cAMP enters the oocyte through gap junctions with the follicle cells (1, 2). Alternatively, cAMP could be generated in the oocyte, and the role of the follicle cells could be to maintain the activity of a stimulatory G protein (Gs) in the oocyte membrane, thus stimulating oocyte adenylyl cyclase (1, 3). Although some evidence has been obtained for each model, neither possibility has been definitively tested.

Figure 1

A method for injecting follicle-enclosed oocytes. (A) Structure of the follicle, also showing an oil drop within the oocyte introduced by microinjection. Drawing modified from (15). (B) Injection chamber and micropipette (4). (C) A follicle containing an injected oocyte. The light spot near the oocyte center is the nucleolus; the light spot near the oocyte periphery is an oil drop introduced by the injection; see (A). (D) Fluorescence image of another follicle containing an injected oocyte; fluorescent dextran in the oocyte cytoplasm indicates a successful injection.

Studies of how meiotic arrest is maintained and released in mammalian oocytes have been limited by a technical problem: the oocyte is embedded in multiple layers of cells that must be left intact to preserve normal regulation. Here, we developed a method for injecting meiotically competent mouse oocytes within antral follicles (260 to 470 μm diameter). The follicle was slightly compressed between two coverslips separated by a 300-μm spacer; this assembly was mounted over a reservoir of medium on a support slide with a coverslip base (Fig. 1B) (4, 5). The chamber was open at the front to allow introduction of a micropipette. The nucleus (germinal vesicle; GV) could be seen in favorable cases, and the nucleolus was always visible (Fig. 1C). The success of the injection was confirmed by including a fluorescent dextran in the injection solution (Fig. 1D).

To examine whether a Gs protein within the oocyte is required to maintain meiotic arrest within the follicle, we microinjected oocytes with an affinity-purified antibody that inhibits Gs function (4, 6–8). This antibody was made against the COOH- terminal 10 amino acids of the α subunit of Gs and specifically recognized both the long and the short forms of Gs (9) in isolated oocytes (Fig. 2A). The amount of Gs protein in oocytes was comparable to that in brain (Fig. 2A), which indicates that Gs is present at a physiologically significant level.

Figure 2

Injection of an antibody against Gscauses oocyte maturation. (A) Immunoblots showing Gs and Gi protein in mouse oocytes and brain probed with the same antibodies used for microinjection. Total protein per lane = 4, 2.5, 5, and 5 μg for lanes 1 to 4, respectively. (B) Follicle-enclosed oocytes were injected with an antibody against Gs or with a control antibody against Gi; the graph shows antibody concentrations in the cytoplasm (the 0 μM point describes uninjected oocytes from follicles processed in parallel). Three hours later, the oocytes were removed from their follicles and scored for the presence or absence of a nucleus (% GVBD). (C) Formation of a polar body by an oocyte injected with the Gs antibody (1.3 μM). The oocyte was removed from the follicle 3 hours after injection, at which time it had undergone GVBD; it was photographed 20 hours later. (D) Control oocyte injected with the Gi antibody (6.7 μM) and removed from the follicle 3 hours later, at which time the GV was intact. (E) Isolated oocytes were injected with antibodies against Gs or Gi, incubated in the presence of 4 mM hypoxanthine (Hx), and scored for GVBD 3 hours later. For (B) and (E), numbers in parentheses indicate the number of oocytes injected with each antibody concentration, and superscript letters indicate the statistical significance of the results compared with Giantibody-injected controls (a, P < 0.03;b, P < 0.01; c,P < 0.001; Fisher's exact test) (4).

The Gs antibody caused resumption of meiosis in follicle-enclosed oocytes, as indicated by light microscopic observation that GV breakdown (GVBD) had occurred in oocytes that were removed from their follicles 3 hours after injection (Fig. 2B). The intracellular concentration of the Gs antibody required to cause GVBD in 50% of the oocytes was between 0.3 and 1.3 μM. Antibody-injected oocytes that underwent GVBD subsequently formed a first polar body (Fig. 2C), which indicates that meiosis progressed normally (14 of 16 oocytes observed).

Control oocytes that were injected with an antibody (6.7 μM) against another G protein, Gi3 (6) (Fig. 2A), and other control oocytes that were not injected were GV intact when removed from their follicles 3 hours after injection (Fig. 2, B and D). The Gi antibody-injected oocytes, as well as control uninjected oocytes, underwent GVBD by 3 hours after removal from the follicle, which indicates that the Gi antibody itself had no detectable effect on oocyte maturation.

These experiments indicated that inhibition of Gs in the oocyte is sufficient to release meiotic arrest and, conversely, that Gs activity in the oocyte is required to maintain meiotic arrest within the follicle. Because the ∼150-kD antibody is too large to pass into follicle cells through gap junctions, which generally have a permeability limit of ∼1 kD (10), the antibody's inhibitory action must be on Gs in the oocyte and not in the follicle cells. It is possible, however, that cAMP entering the oocyte through gap junctions also contributes to the maintenance of meiotic arrest.

Oocytes that are removed from their follicles and maintained in a cAMP phosphodiesterase inhibitor such as hypoxanthine stay arrested at meiotic prophase (1, 3, 4), which suggests that isolated oocytes might have enough Gsactivity to keep cAMP elevated if its hydrolysis is prevented. To examine this, we injected isolated oocytes with the Gsantibody and cultured them in the presence of hypoxanthine. GVBD occurred in these oocytes but not in controls injected with the Gi antibody (Fig. 2E), which indicates that, even in the absence of an intact follicle, at least some Gs activity remains in the oocyte (4). Additional controls showed that the Gs antibody did not cause GVBD by an action downstream of cAMP and that Gs antibody-matured oocytes responded normally to fertilization (4) (fig. S1).

Gs alone has little constitutive activity (11), which suggests that Gs activity in the oocyte might be maintained by a Gs-linked receptor in the oocyte membrane. Because G-protein–linked receptors can have high constitutive activity (12, 13), this could potentially account for the Gs activity in isolated oocytes. In the intact follicle, a ligand might activate such a receptor further, possibly explaining the maintenance of meiotic arrest by the follicular environment (4). It is also possible that Gs activity in the oocyte is the same regardless of whether the follicle is present and that the meiosis-inhibiting signal from the follicle acts at another level of the cAMP pathway—for example, to suppress cAMP phosphodiesterase activity in the oocyte (1, 14).

Comparing these findings with previous studies of frog (Xenopus laevis) oocytes, in which the same Gsantibody was found to cause GVBD (8), indicates that Gs activity may be a conserved mechanism for maintaining meiotic arrest in vertebrate oocytes (4) (fig. S2). A difference between mouse and Xenopus, however, is thatXenopus oocytes do not undergo spontaneous GVBD when removed from their follicles. This could be due to species differences in the level of constitutive activity of Gs-linked receptors in the oocyte membrane or of phosphodiesterase activity in the oocyte. By identifying oocyte Gs as a required factor in maintaining meiotic arrest, our results open the way for future investigations of signaling molecules from the follicle that maintain meiotic arrest in the oocyte as well as the mechanism by which luteinizing hormone causes the resumption of meiosis.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 and S2


  • * To whom correspondence should be addressed. E-mail: lmehlman{at}


View Abstract

Navigate This Article