Microinjection of a 19-kDa guanine nucleotide-binding protein inhibits maturation of Xenopus oocytes.

ADP-ribosylation factors (ARFs) are 19-21-kDa proteins purified from bovine brain that bind guanosine 5'-triphosphate (GTP). They exhibit GTP-dependent activity as activators of cholera toxin-catalyzed ADP-ribosylation of the alpha-subunit of the stimulatory guanine nucleotide-binding protein of the adenylyl cyclase system (Gs alpha). ARF, which interacts directly with the catalytic subunit of cholera toxin, has no known physiologic role. Intracellular microinjection of ARF was employed to investigate the effect of ARF on progesterone- and insulin-stimulated maturation of Xenopus oocytes. Maturation was inhibited by injection of ARF 3-8 h before exposure of oocytes to progesterone or insulin. ARF inhibition was dependent on progesterone concentration but not on insulin concentration. Inhibition was enhanced by concomitant injection of GTP and to a greater extent by guanosine 5'-O-(thiotriphosphate) (GTP gamma S) which, in the absence of ARF, inhibited somewhat at early time points. The demonstration of this effect of ARF on both progesterone- and insulin-stimulated oocyte maturation may provide a clue to the physiologic role of this guanine nucleotide-binding protein.

Both of these proteins bind GTP (2,8,9), and ras p21 exhibits GTPase activity (8,9) which has thus far not been detected with ARF (2). It has been shown that injection of ras p21 into Xenopus oocytes induces maturation (IO), whereas injection of monoclonal antibody against ras p21 prevents insulin-but not progesterone-induced maturation (11,12). T o determine whether ARF might have functional effects like those of ras p21 in Xenopus oocytes, the studies reported here were initiated.

EXPERIMENTAL PROCEDURES
Oocytes were harvested from 2-2.5-year-old virgin female Xenopus laevis toads (Nasco, Inc., Fort Atkinston, WI) using tricaine anesthesia, 3-5 days after injection of 25-35 IU human chorionic gonadotropin (Sigma) into the dorsal lymph sack. After incubation in calcium-free OR-2 (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgC12, 5 mM Hepes, pH 7.6) containing collagenase (Sigma, type lA), 2.5 mg/ml, for 5-7 h at 20 "C, oocytes to be stimulated by progesterone were washed twice a t 20 "c in ND-96 (96 mM NaC1, 2.0 mM KCl, 1.8 mM CaC12, 1 mM MgC12,5 mM Hepes, pH 7.6) containing 2.5 mM pyruvate, penicillin (100 units/ml), and streptomycin (100 pg/ml), and incubated overnight in the same medium. Oocytes to be stimulated by insulin were washed twice and incubated overnight in modified OR-2 (83 mM NaCI, 2.5 mM KCI, 1 mM Cac12, 1 mM MgC12, 10 mM Hepes, pH 7.6) with bovine serum albumin (fraction V, Miles Laboratories, Elkhart, IN), 1 mg/ml, and antibiotics and pyruvate as above. They were placed in modified ND-96 (96 mM NaC1, 0.5 mM CaC12, 1 mM MgC12, 10 mM Hepes, pH 7.8) with bovine serum albumin, 5 mg/ml, and penicillin, streptomycin, and pyruvate as above, 1-2 h before injection. Stage VI (see Ref. 13) oocytes were selected, transferred to plastic dishes containing 6 ml of the appropriate medium (-250 oocytes/dish), and kept overnight at 20 "C with gentle agitation. Next morning, damaged oocytes were discarded, and oocytes, with or without injection as indicated for each experiment, were distributed to plastic dishes containing 6 ml of medium. Progesterone (Sigma) or insulin (regular Iletin I, U-100, Lilly) was added as described for each experiment, and at the indicated times thereafter oocytes were inspected to evaluate appearance of the maturation spot, signifying germinal vesicle breakdown (GVBD). The number of oocytes with GVBD was recorded and percent GVBD calculated. 00cytes with equivocal morphologic changes at the end of each experiment were fixed in 10% trichloroacetic acid and dissected to verify the status of the germinal vesicle. Progesterone solutions were prepared from ethanol stock; ethanol concentration in the incubation medium was 0.05%. Insulin was added directly to the incubation medium to yield the indicated final concentration. A pressure injection system and micropipette with outside tip diameter of 12-15 pm were used for microinjection. Solutions of injected proteins and nucleotides were prepared in 10 mM NaCl, 10 mM Hepes, pH 7.5.
ARF was purified from bovine brain as described for the preparation of sARFII in Ref. 3. Where indicated, proteins were incubated with guanine nucleotide for 50 min at 20 "C before injection. For calculation of concentrations of ARF solutions a M, of 19,000 was assumed. Injected volume was 50 nl for progesterone-and 35-40 nl for insulin-treated oocytes. If injected ARFs were uniformly distributed in total cell water (approximately 500 nl/oocyte), intracellular concentrations would be 7-10% of those injected. Data are means plus or minus standard error of the mean for values from two or more replicate dishes containing the indicated number of oocytes. When multiple comparisons between treatment groups were made, one-way analysis of variance and the Newman-Keuls procedure for multiple comparison were used at each time point in progesterone-stimulated oocytes, The Tukey procedure for multiple comparisons was used in insulin-stimulated oocytes. Unpaired t tests were used otherwise.

RESULTS AND DISCUSSION
When oocyte maturation, as determined by GVBD, was induced by addition of 10 nM progesterone to the incubation medium, maturation of a group of oocytes was usually complete within 7 h (Fig. l), although there was some variation between oocytes harvested from different toads and from the same toad at different times. Concentrations of progesterone greater than 10 nM caused more rapid maturation, and concentrations less than 10 nM failed to induce maturation of 90-100% of oocytes in 7 h (Fig. 1). Injection of ARF into oocytes 6 h before addition of 10-50 nM progesterone inhibited GVBD, and inhibition was dependent on the concentration of injected ARF (Fig. 2). Inhibition was evident as early as 4 h after addition of progesterone and persisted for the duration of the experiment although diminishing with time (18 h). At all observation times, inhibition of GVBD in oocytes injected with 10, 20, or 30 p~ ARF (9.5, 19, and 28 ng/oocyte, respectively) was significant ( p < 0.001); 3 p M ARF had no measurable effect (Fig. 2). Inhibition of GVBD produced by 10 p~ ARF was less than that produced by 20 or 30 p~ ARF at times between 5 and 18 h ( p < 0.001) as well as at 4 h ( p < 0.05) after progesterone was added; 20 p M ARF was less inhibitory than 30 p~ ARF a t 7 ( p < 0.06), 8, and 18 h ( p < 0.001) after progesterone.
ARF inhibition was dependent on the concentration of progesterone used to stimulate GVBD and on the time elapsed between ARF injection and addition of progesterone (Fig. 3). GVBD stimulated by 20 nM progesterone was only slightly inhibited by 30 p~ ARF whether injected 1 or 5 h before stimulation, and inhibition was negligible when ARF was injected 12 or 15 h before stimulation (Fig. 3). Significant inhibition was produced by injection of 30 p~ ARF, 3 ( p < 0.03), or 8 ( p < 0.006), but not 15 h before addition of 10 nM progesterone. These findings would be consistent with the  conclusion that either active ARF itself or responsiveness to ARF does not persist 15 h after injection.
With oocytes harvested from the same toads at a different time of year, 24 p~ ARF caused 75% inhibition of GVBD stimulated by 50 nM progesterone ( p < 0.002). It appears that there is a seasonal variation in oocyte sensitivity to progesterone and the greatest inhibitory effect of ARF is seen when the concentration of progesterone used to stimulate maturation is near the minimum necessary to cause a full response of 90-100% GVBD within 7 h (Fig. 1).
As ARF activity (when assessed by its ability to activate cholera toxin) requires the presence of GTP or a GTP analogue (4), the effects of guanine nucleotides on GVBD were investigated (Fig. 4). Injection of ovalbumin alone or in combination with GDP or GTP had no effect on progesteronestimulated GVBD, nor did injection of boiled ARF. G T P r S (plus ovalbumin) produced slight but significant inhibition of GVBD at 4,5, 6 ( p < 0.001), and 7 h ( p < 0.05) after addition of progesterone, but not at 8 nor 18 h. Inhibition of GVBD by ARF was not altered by concomitant injection of GDP (Fig. 4). ARF inhibition (with or without GDP) was significant ( p < 0.001) at all times compared to ovalbumin with or without GDP o r GTP, or boiled ARF, and ranged from 32 to 79%. Inhibition was not significant compared to ovalbumin plus G T P r S a t 4 and 5 h. Coinjection of GTP significantly enhanced ARF inhibition a t 5, 6, 7 ( p < 0.001), 8, and 18 h ( p < 0.05). Injection of ARF plus GTPyS caused virtually complete inhibition of GVBD for a t least 18 h (Fig. 4). Inhibition was greater than that seen in any other group at 5 , 6, 7, 8, and 18 h ( p < 0.001) (except inhibition compared to ARF plus GTP was not significant at 5 h).
To determine whether the action of ARF could be localized to a progesterone-specific portion of the pathway leading to oocyte maturation, its effects on GVBD induced by insulin were investigated. For these studies, the medium used for progesterone experiments was modified to improve responsiveness to insulin by increasing pH, omitting KCl, and adding bovine serum albumin. Insulin concentrations between 1 and 7 ~L M were needed to cause maximal maturation of oocytes, which occurred somewhat more slowly than did progesteronestimulated maturation (Fig. 5). With insulin-stimulated oocytes, injection of 24 p~ ARF (19 ng/oocyte) resulted in inhibition as early as 7-9 h after hormone addition (Fig. 5 ) that persisted for 36 h (data not shown). As observed with progesterone-stimulated maturation, percentage inhibition tended to diminish with time. The inhibitory effect of ARF was not dependent on insulin concentration. At 7 h, inhibition of GVBD was 59% ( p < 0.03) with 7 PM insulin, and 52% ( p < 0.02) with 1 ~L M insulin. In this regard (as in some others) insulin stimulation of maturation appears to differ from progesterone stimulation, since ARF inhibition of the latter was best demonstrated with low progesterone concentrations. It is possible that the inhibitory effect of ARF on GVBD induced by these relatively low concentrations of progesterone (which were nonetheless sufficient to cause a brisk and complete oocyte response) could result from an ARF-induced increase in progesterone uptake or metabolism. Since ARF also inhibits insulin-stimulated GVBD, however, its effects on oocyte maturation are presumably not due solely to increased removal or inactivation of progesterone. It seems more likely that ARF inhibits both progesterone-and insulin-induced GVBD by acting a t a site in the maturation pathway that is common to both.
When insulin was used as the stimulating hormone, GTPyS (plus ovalbumin) was inhibitory at 9.0 and 10.5 h, but inhibition was not significant a t later times (Fig. 5 ) . G T P (plus ovalbumin) did not cause significant inhibition of GVBD at any time. Inhibition by ARF (or ARF plus GTP, which was not different) was significant (compared to ovalbumin with our without GTP) at all times after 6.25 h with inhibition ranging from 45 to 100% ( p < 0.05). With ARF plus GTP+, inhibition was greater than in any other ARF-injected group at 13.25 and 16.5 h ( p < 0.05). Inhibition of GVBD in ARFinjected oocytes decreased with time; however, 36 h after insulin addition, inhibition by ARF or ARF plus G T P r S (relative to oocytes injected with ovalbumin with or without nucleotide) was still 36-46% ( p < 0.05).
Whereas injection of 9.5-19 ng of ARF/oocyte, as described above, caused significant inhibition of hormone-stimulated maturation, injection of ARF (or ovalbumin) had no discernible effects for a t least 72 h on oocytes not exposed to hormone (data not shown). These nanogram amounts of ARF represent less than 0.02% of the total cellular protein, assumed to be approximately 280 pg/oocyte (14). Higher concentrations of ras p21 (10-50 ng/oocyte) have been reported to stimulate oocyte maturation (10)(11)(12)15). ARF appears to have an effect opposite to that described for Ha-ras p21. This may suggest that two distinct classes of low molecular weight GTP-binding proteins exist, each with opposing effects. It is not clear whether other endogenous ARF-or ras-like GTP-binding proteins may affect oocyte maturation; conceivably, ARF or ras is mimicking the action of other GTP-binding proteins.
In uninjected oocytes, a small amount of immunoreactive ARF was detected in the soluble fraction by immunoblot analysis using anti-ARF polyclonal antibodies (Fig. 6). This assay was capable of detecting 25-50 ng of bovine ARF/lane. After ARF injection, additional soluble immunoreactive ARF was apparent, forming a doublet with endogenous ARF. This suggests that endogenous ARF in the oocyte has electrophoretic mobility similar to that of bovine sARFI (3) which differs slightly from sARFII. Immunoreactive ARF seen immediately following injection of 50 nl of 30 p~ ARF/oocyte remained present through 15 h (Fig. 6). No ARF was detected in membrane fractions of oocytes either before or after injection (data not shown). The failure of ARF to inhibit GVBD when injected 15 h before progesterone stimulation and the decrease in inhibition of both progesterone-and insulin-stimulated maturation by ARF with time may be due to the disappearance of an active form with no change in immunoreactive ARF or to desensitization of the oocyte, although alternative explanations have not been excluded.
The heterotrimeric guanine nucleotide-binding proteins are components of several transmembrane signaling systems, linking cell surface receptors with their intracellular effectors. These so-called "G-proteins" are functionally regulated by guanine nucleotides and possess GTPase activity, GTP binding causes activation that is terminated by GTP hydrolysis, and nonhydrolyzable GTP analogues such as GTPyS cause persistent activation (see Ref. 16 for review). Similarly, ras p21 binds GTP, has GTPase activity, and its function is regulated by GTP binding and hydrolysis (8,9,17,18). Enhancement of cholera toxin-catalyzed ADP-ribosylation by ARF requires GTP or a GTP analogue (2)(3)(4). The finding that ARF inhibition of hormone-stimulated oocyte maturation is maximally enhanced by GTPyS may suggest that this ARF effect is likewise regulated by guanine nucleotide binding, although parallel or synergistic effects of ARF and GTPyS could be postulated. ARF has no detectable GTPase activity (2). However, purified ARF contains bound GDP (2), and the nonhydrolyzable analogue GTPyS was more effective than GTP in enhancing ARF inhibition of GVBD. Both findings are consistent with the possibility that ARF GTPase activity may be demonstrable under suitable conditions. In the case of ras, for example, hydrolysis of GTP by ras p21 is markedly increased by cytosol from Xenopus oocytes, and a protein that activates the ras p21 GTPase has recently been purified from bovine brain (19,20).

14.3-
Late stage oocytes are normally arrested in the first meiotic prophase. After stimulation with hormones such as insulin (21,22) or progesterone (23) that act on cell surface receptors (24, 25), meiosis is reinitiated and the oocyte proceeds to the second meiotic metaphase with concurrent disintegration of the nuclear membrane (GVBD) (21, 26, 27). This process is accompanied by a receptor-mediated increase in calcium flux which may not be a necessary concomitant of maturation (28), an increase in intracellular pH, a decrease in cAMP content resulting from changes in adenylyl cyclase and phosphodiesterase activities (29)(30)(31), changes in the intracellular concentration of maturation promoting factor (32, 33), and changes in the phosphorylation of ribosomal S6 protein and other proteins (34-39) catalyzed by multiple kinases including a serine-specific S6 protein kinase that seems itself to be regulated by phosphorylation (40). Both the adenylyl cyclase and phosphoinositide signaling systems have been implicated in the control of oocyte maturation. A decrease in cAMP is sufficient to cause oocyte maturation, and stimulation of CAMP-dependent protein kinase is sufficient to inhibit maturation (41). Receptor-mediated changes in calcium, however, may induce or accelerate oocyte maturation independent of small changes in cAMP levels (42), and the phosphoinositide signaling system may be active during insulin-induced maturation as suggested by the permissive or stimulatory effects of inositol 3-phosphate and a phorbol ester on oocyte maturation (43).
Several different mechanisms of action for ras p21 on Xenopus oocytes have been postulated. Antibodies against ras p21 enhance progesterone-stimulated GVBD, possibly due to inhibition of adenylyl cyclase (44). It has also been reported that ras p21 can induce maturation with no apparent effect on cAMP levels (10). A recent report suggests that Ha-ras p21 injection may activate a pool of oocyte cAMP phosphodiesterase that is normally stimulated by insulin (15). Furthermore, Ha-ras p21 may act via stimulation of diacylglycerol production (45).
The mechanism of ARF action on Xenopus oocyte maturation is unclear. There are potentially numerous sites at which ARF might act to interfere with both progesteroneand insulin-stimulated oocyte maturation. ARF is defined by its ability to stimulate the ADP-ribosyltransferase activity of cholera toxin, suggesting the possibility that ARF functions in animal cells to enhance the activity of an endogenous ADPribosyltransferase(s). Thus far, however, no effects of ARF on four extensively purified eukaryotic ADP-ribosyltransferases have been observed.2 It appears that the insulin-and progesterone-stimulated pathways may converge at the level of the second messenger cAMP and share subsequent steps (e.g. maturation promoting factor activation, protein phosphorylation). ARF might affect maturation at one of these sites or might interfere with some basic cellular process such as cytoskeletal function or vesicular transport. Whether the physiologic target of ARF is structurally or functionally related to cholera toxin remains to be determined. Additional research is needed to define further the effects of ARF on the oocyte as well as on other types of cells, and to identify cellular enzymes or processes that are influenced by this guanine nucleotide-binding protein.