Reports

The Biochemical Basis of an All-or-None Cell Fate Switch in Xenopus Oocytes

See allHide authors and affiliations

Science  08 May 1998:
Vol. 280, Issue 5365, pp. 895-898
DOI: 10.1126/science.280.5365.895

Abstract

Xenopus oocytes convert a continuously variable stimulus, the concentration of the maturation-inducing hormone progesterone, into an all-or-none biological response—oocyte maturation. Here evidence is presented that the all-or-none character of the response is generated by the mitogen-activated protein kinase (MAPK) cascade. Analysis of individual oocytes showed that the response of MAPK to progesterone or Mos was equivalent to that of a cooperative enzyme with a Hill coefficient of at least 35, more than 10 times the Hill coefficient for the binding of oxygen to hemoglobin. The response can be accounted for by the intrinsic ultrasensitivity of the oocyte's MAPK cascade and a positive feedback loop in which the cascade is embedded. These findings provide a biochemical rationale for the all-or-none character of this cell fate switch.

Fully grown Xenopus laevis oocytes are arrested in a state that resembles the G2 phase of the cell division cycle with inactive cyclin-dependent kinase Cdc2 and an intact germinal vesicle. Exposure to the hormone progesterone induces oocytes to undergo maturation, during which they activate Cdc2, undergo germinal vesicle breakdown, complete the first meiotic division, and finally arrest in metaphase of meiosis 2 (1). Oocyte maturation is an example of a true cell fate switch; oocytes can reside in either the G2arrest or the metaphase arrest state for extended periods of time, but can be in intermediate states only transiently.

Progesterone-induced maturation is thought to be triggered by activation of a cascade of protein kinases—Mos, Mek-1, and p42 or Erk2 MAP kinase (MAPK). Progesterone causes the accumulation of Mos, which phosphorylates and activates Mek-1. Active Mek-1 in turn phosphorylates and activates p42 MAPK, which brings about activation of the Cdc2–cyclin B complex. Interfering with the accumulation of Mos (2) or the activation of Mek-1 (3) or p42 MAPK (4) inhibits progesterone-induced activation of Cdc2 and maturation, and microinjection of nondegradable Mos (5), constitutively active Mek-1 (4, 6), or thiophosphorylated p42 MAPK (7) brings about Cdc2 activation and maturation in the absence of progesterone. At some point in this chain of events, a continuously variable stimulus—the progesterone concentration—is converted into an all-or-none biological response.

Studies of the steady-state responses of the MAPK cascade inXenopus oocyte extracts indicate that the cascade might contribute to the all-or-none character of oocyte maturation. In extracts, the response of MAPK to recombinant malE-Mos (a maltose-binding protein Mos fusion protein) is highly ultrasensitive (8), meaning it resembles that of a positively cooperative enzyme (9). The apparent Hill coefficient (n H, a measure of the ultrasensitivity) for the MAPK response is about 5 (8). This is a large Hill coefficient; the benchmark is the Hill coefficient for oxygen binding by hemoglobin, which is about 2.8 (10). The ultrasensitivity arises in part from the fact that MAPK requires the phosphorylation of two sites for activation (11,12), and it increases nearly multiplicatively as the cascade is descended (13). An ultrasensitive system behaves more like a switch than a Michaelian (n H = 1) system does—the response to small stimuli is minimal, but once the system begins to respond, it switches from off to on over a narrower range of stimulus concentrations than does a Michaelian system (Fig.1G). Thus, the MAPK cascade might contribute to the all-or-none character of oocyte maturation, provided ultrasensitivity is exhibited by MAPK in intact oocytes as well as extracts.

Figure 1

Responses of oocytes to progesterone. (A) Overall responses. Each point represents a sample of 11 to 39 oocytes. Error bars denote two standard errors of the mean. MAPK-P, phosphorylated MAPK. (B andC) Two possible origins of a graded response. (D) Calculated distributions of oocytes incubated with a half-maximal stimulus for various assumed values of the Hill coefficient (n H) for the individual oocytes' responses. The oocyte-to-oocyte variability was assumed here to correspond to an m = 1 curve (15). (E) MAPK immunoblots for individual oocytes treated with progesterone. The first two lanes of each blot represent oocytes treated with no progesterone (–) or 8 μM progesterone (+). (F) Cumulative MAPK phosphorylation data fromN = 209 individual oocytes. (G) Stimulus-response curves inferred for p42 MAPK phosphorylation in vivo (n H ≈ 42) and measured for p42 MAPK activation in vitro [n H ≈ 5, from (8)]. A Michaelian (n H = 1) curve is shown for comparison.

We, therefore, assessed the phosphorylation of p42 MAPK in groups of oocytes treated with different concentrations of progesterone (14). The overall response appeared to be no more switchlike than that of a typical Michaelian system (n H ≈ 1, Fig. 1A). However, a problem arises in interpreting the response of a potentially heterogeneous population. A graded overall response could mean that each of the individual oocytes had a graded response (Fig. 1B), but even if individual oocytes had perfectly switchlike responses, samples of oocytes would yield a graded response if the oocytes varied with respect to the concentration of progesterone required to switch them on (Fig. 1C).

These two possibilities can be distinguished by examining individual oocytes treated with intermediate concentrations of progesterone. If the individual responses are graded, each oocyte should have an intermediate amount of MAPK phosphorylation (Fig. 1B); if they are switchlike, the oocytes should have either very high or very low levels of MAPK phosphorylation (Fig. 1C). This argument can be translated into a mathematical formula (15) for inferring the steepness (value of n H) of the oocytes' individual responses from the observed distribution of responses in a sample of oocytes (Fig. 1D).

Accordingly, we examined the steady-state phosphorylation of MAPK in 190 individual progesterone-treated oocytes and 19 individual untreated oocytes. Every oocyte had either very high (>90% of maximal) or very low (<10% of maximal) amounts of MAPK phosphorylation (Fig. 1, E and F). Thus, the response of the individual oocytes was essentially all-or-none; a lower bound for the Hill coefficient was calculated to be 42 (15, 16) (Fig. 1G).

To determine whether the all-or-none character of the response was generated by the MAPK cascade, or was passed down to the cascade by upstream signaling elements, we microinjected oocytes with purified malE-Mos, a direct activator of Mek-1 (17), and assessed the resulting MAPK phosphorylation. The response of the population was steep (n H > 5, Fig.2A), and only one Mos-injected oocyte was found with an intermediate amount of MAPK phosphorylation (out of 89; Fig. 2, C and D). The Hill coefficient inferred for the oocytes' individual responses was about 35, similar to that for the response to progesterone. Thus, the MAPK cascade can generate, not simply propagate, a highly switchlike response to a continuously variable stimulus.

Figure 2

Responses of oocytes to microinjected malE-Mos in the presence and absence of protein synthesis. The responses of pools of 10 to 20 oocytes to microinjection of malE-Mos in the absence (A) and presence (B) of cycloheximide (10 μg/ml). Error bars denote two standard errors of the mean. (C) MAPK immunoblots for individual oocytes microinjected with malE-Mos. The first two lanes of each blot represent oocytes injected with water (–) or 200 nM malE-Mos (+). (D) Cumulative data from 202 individual oocytes.

The responses seen in intact oocytes (Figs. 1 and 2) were much more switchlike than those seen in oocyte extracts (8). We hypothesized that a key difference could be a positive feedback loop known to operate in intact oocytes (4, 18-20), and known not to operate in extracts (8), whereby MAPK or something downstream from MAPK promotes the stabilization and accumulation of Mos, at least in part through the phosphorylation of Ser3 (Fig. 3A) (19). A positive feedback loop would markedly increase the abruptness of the MAPK cascade's response (Fig. 3B) (21).

Figure 3

Ultrasensitivity as a means of providing a positive feedback system with both a stable off state and a stable on state. (A) Mos synthesis, destruction, and stabilization reactions. The feedback from MAPK to Mos could be direct (18), as shown here, or indirect (20). (B) Calculated stimulus-response curves for MAPK phosphorylation assuming no ultrasensitivity or feedback (denoted “neither”), ultrasensitivity but no feedback, feedback but no ultrasensitivity, or feedback plus ultrasensitivity (denoted “both”). Curves were calculated as described (21) with values of n H, n H′, EC50, and EC50′ estimated from studies in extracts (8) and the present studies (n H ≈ 5, n H′ ≈ 3, EC50 ≈ 27 nM, and EC50′ ≈ 20 nM), and takingk 1 k 2/k −1 k −2to be 0.5. The curves shown assume the system was initially in its off state. Graphical depiction of the expected steady-state level of Mos-P with positive feedback and either (C) a Michaelian response from the MAPK cascade or (D) an ultrasensitive response from the MAPK cascade. Stable steady states are denoted ss. Unstable steady states are denoted by asterisks. The arrows show in which direction the system would be driven if perturbed from a steady state.

If protein synthesis–dependent positive feedback contributed to the highly switchlike responses seen in intact oocytes, then the protein synthesis inhibitor cycloheximide should make the response of oocytes to malE-Mos more like that seen in extracts. In agreement with this prediction, a large proportion of cycloheximide-treated, Mos-injected oocytes had intermediate amounts of MAPK phosphorylation (Fig. 2, C and D). These results imply a Hill coefficient of about 3, similar to that seen in extracts (22). Thus, the MAPK cascade does exhibit some ultrasensitivity even when positive feedback is precluded, but protein synthesis allows a more highly switchlike response.

The intrinsic ultrasensitivity of the MAPK cascade and the protein synthesis–dependent positive feedback loop together should produce a more satisfactory switch than either mechanism alone would. This can be seen through quantitative modeling (Fig. 3B) or simple graphical arguments (Fig. 3, C and D). If positive feedback operated and the MAPK cascade exhibited a Michaelian response to Mos, then the system would have a stable on state and an unstable off state (Fig. 3, B and C). Any nonzero level of Mos phosphorylation, added malE-Mos, or MAPK activity would trigger the feedback loop and drive the system to its on state. However, if the MAPK cascade exhibited an ultrasensitive response to Mos, then the system would have both a stable off state and a stable on state separated by a threshold (Fig.3, B and D) (23). The ultrasensitivity of the MAPK cascade essentially filters small stimuli out of the feedback loop.

In summary, the MAPK cascade is activated in a highly ultrasensitive—essentially all-or-none—fashion duringXenopus oocyte maturation. This behavior is proposed to arise from two known properties of the oocyte's MAPK cascade: positive feedback, which ensures that the occyte cannot rest in a state with intermediate MAPK phosphorylation, and the cascade's intrinsic ultrasensitivity, which establishes a threshold for activation of the positive feedback loop. Positive feedback does not appear to be uncommon, and there are many mechanisms that can give rise to ultrasensitivity (8, 9, 11, 24). Thus, other biological switches may be constructed from components that are similar or analogous to those used by the oocyte.

  • * To whom correspondence should be addressed. E-mail: ferrell{at}cmgm.stanford.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article