Mammalian oocytes grow within ovarian follicles in which the oocyte is coupled to surrounding granulosa cells by gap junctions. We report here that growing oocytes isolated from mouse preantral follicles are incapable of recovering from an experimentally induced acidosis, and that oocytes acquire the ability to manage acid loads by activating Na+/H+exchange during growth. By contrast, granulosa cells from similar preantral follicles possess substantial Na+/H+ exchange capacity,which is attributable to the simultaneous action of two Na+/H+ exchanger isoforms: NHE1 and NHE3. Granulosa cells were also found to possess a V-type H+-ATPase that drives partial acidosis recovery when Na+/H+ exchange is inactivated. By monitoring intracellular pH (pHi) in small follicle-enclosed oocytes, we found that the oocyte has access to each of these acidosis-correcting activities, such that small follicle-enclosed oocytes readily recover from acidosis in a manner resembling granulosa cells. However, follicle-enclosed oocytes are unable to access these activities if gap-junction communication within the follicle is inhibited. Together, these experiments identify the NHE isoforms involved in regulating oocyte pHi, indicate that gap junctions allow granulosa cells to exogenously regulate oocyte pHi against acidosis until the oocyte has acquired endogenous pHi regulation, and reveal that granulosa cells possess multiple mechanisms for carrying out this function.
Regulation against unwelcome perturbations in intracellular pH(pHi) is a fundamental and essential process carried out by virtually all cells. The principal pHi regulatory mechanisms in mammalian cells are HCO3-/Cl- exchangers of the AE (anion exchange) family, which export HCO3- in exchange for Cl- thereby correcting pHi increases(Romero et al., 2004; Alper, 1994), and Na+/H+ exchangers of the NHE (also known as Scl9a - Mouse Genome Informatics) gene family which correct pHi decreases by exporting protons in exchange for Na+(Orlowski and Grinstein, 1997; Orlowski and Grinstein, 2004). Multiple isoforms of the Na+/H+ exchanger (NHE1-9) and the HCO3-/Cl- exchanger (AE1-3, with numerous splice variants) exist, with heterogeneous tissue expression that probably reflects their different kinetic properties and regulation(Romero et al., 2004; Orlowski and Grinstein, 2004). These transport systems may be supplemented by additional components, such as Na+, HCO3-/Cl- exchangers (NBC),which relieve acidosis in some cell types by importing HCO3- (Romero et al., 2004), and V-type H+-ATPases (V-ATPases) which export protons across the plasmalemma in some cells(Nelson and Harvey, 1999; Merzendorfer et al., 1997; Kawasaki-Nishi et al., 2003). The multifaceted array of pHi regulatory mechanisms provides the basis for tight control of intracellular pH in the face of external pH changes and metabolic acid generation. Perhaps unsurprisingly, impairment of pHi regulation can compromise cell function and viability. For example, growth and proliferation is impaired in some pHiregulation-compromised cells when intracellular pHi is disturbed(Grinstein et al., 1989; Kapus et al., 1994), and pHi dysregulation jeopardizes cell survival(Pouyssegur et al., 1984). pHi regulation is of particular importance in early mammalian development, as inhibition of pHi regulatory mechanisms hinders preimplantation development in mouse and hamster embryos(Lane et al., 1998; Zhao et al., 1995).
Mammalian oocytes grow within ovarian follicles, discrete micro-organs consisting of the oocyte and a surrounding shell of (somatic) granulosa cells. Oocyte growth takes about 15 days in mouse, during which time the oocyte increases in diameter from about 15 to 70-80 μm. Throughout growth, the oocyte remains coupled to its granulosa cells by gap junctions, intercellular channels that permit small molecules (<∼1 kDa) to pass freely between apposing cells (Anderson and Albertini,1976; Ducibella et al.,1975; Kidder and Mhawi,2002). Growth of the oocyte is absolutely dependent upon its association with granulosa cells (Eppig,1977; Eppig, 1979; Brower and Shultz, 1982; Eppig and Wigglesworth, 2000),and the importance of gap junction communication is well established, since oocyte growth is prevented by targeted deletion of gap junctions found within the follicle (Ackert et al.,2001; Simon et al.,1997). Gap junctions provide a means for granulosa cells to assist the oocyte in carrying out metabolic functions that the growing oocyte is not yet capable of. For example, granulosa cells promote oocyte uptake of some amino acids and nucleotides that are ineffectively taken up by denuded oocytes(Colonna and Mangia, 1983; Cross and Brinster, 1974; Haghighat and van Winkle,1990; Heller et al.,1981; Heller and Schultz,1980). In addition, although oocytes metabolise glucose poorly,granulosa cells metabolise glucose efficiently, and provide the oocyte with metabolic substrates that it can use(Biggers et al., 1967). However, although the ability of the granulosa cells to provide the oocyte with substrates is well established, there has been little evidence that granulosa cells assume homeostatic functions on behalf of the oocyte.
We recently showed that smaller growing oocytes isolated from the ovary with their granulosa cells removed (`denuded oocytes'), had a very limited ability to recover from an induced alkalosis, whereas fully grown denuded oocytes recovered robustly (Erdogan et al., 2005). The ability to recover from alkalosis was acquired during oocyte growth as a result of activation of HCO3-/Cl- exchange, which was inactive in very small oocytes (∼20 μm), showed only limited activity in denuded oocytes up to about 60 μm in diameter, and then became fully activated in oocytes when they were nearly fully grown (>65 μm) and competent to complete meiosis. Using an assay in which any Na+/H+exchanger activity could be revealed by an amiloride-sensitive intracellular acidification upon external Na+ removal, there was also evidence that Na+/H+ exchanger activity follows a similar pattern during oocyte growth (Erdogan et al.,2005). Thus, we concluded that small growing oocytes do not possess intrinsic pHi regulatory mechanisms, but develop them when they are nearly fully grown.
Subsequently it was found that, in the intact follicle, oocytes have access to significant amounts of alkalosis-correcting HCO3-/Cl- exchange enabling them to recover from alkalosis, and that granulosa cells surrounding growing oocytes had substantial HCO3-/Cl- exchange, which led us to propose that ovarian oocytes may be able to access HCO3-/Cl- exchangers in the granulosa cells(FitzHarris and Baltz, 2006). However, it was difficult to discount the possibility that granulosa cells may activate pHi regulation endogenous to the oocyte(FitzHarris and Baltz, 2006),and the lack of isoform-specific HCO3-/Cl-exchange inhibitors prevented identification of the specific alkalosis-correcting mechanisms present in granulosa cells and fully grown oocytes.
The ability of growing oocytes to regulate against acidosis has not yet been addressed, nor have the molecular identities of any pHi-regulatory mechanisms in oocytes or surrounding granulosa cells been determined. Our experiments described here indicate that growing oocytes rely upon granulosa cells to regulate against acidosis until the oocyte's own Na+/H+ activity is activated during growth, and provide strong evidence that ooplasmic pHi is regulated by pHi-regulatory mechanisms within granulosa cells, which possess multiple mechanisms that can raise oocyte pHi, including two Na+/H+ exchanger isoforms and a proton-extruding V-ATPase, some of which are not present even in the fully grown oocyte.
MATERIALS AND METHODS
Chemicals and solutions
Chemicals and drugs were obtained from Sigma (St Louis, MO, USA) unless otherwise noted. Carboxyseminaphthorhodafluor-1-acetoxymethyl ester(SNARF-1-AM), and SNARF-1-dextran were obtained from Molecular Probes (Eugene,OR, USA). Cariporide and S3226 were kindly supplied by Aventis. Stock solutions were prepared in either water [dibutyrladenosine 3,5-cyclic monophosphate (dbcAMP)], ethanol (nigericin, NEM) or dimethyl sulfoxide (DMSO;SNARF-1-AM, valinomycin, DIDS, 1-octanol, 18α-glycyrhetinic acid,cariporide, S3226, concanamycin).
All media were based on KSOM mouse embryo culture medium(Lawitts and Biggers, 1993)which contains (in mM) 95 NaCl, 2.5 KCl, 0.35 KH2PO4,0.2 MgSO4, 10 sodium lactate, 0.2 glucose, 0.2 sodium pyruvate, 25 NaHCO3, 1.7 CaCl2, 1 glutamine, 0.01 tetra sodium EDTA,0.03 streptomycin sulphate and 0.16 penicillin G, and 1 mg/ml bovine serum albumin (BSA). Hepes-KSOM was used for oocyte collection and microinjection(21 mM Hepes replacing equimolar NaHCO3, pH adjusted to 7.4). For fluorophore-loading and pHi measurements, sodium lactate was reduced to 1 mM. Where used, bicarbonate-free medium was prepared by replacing NaHCO3 with equimolar NaCl. Experiments were performed in bicarbonate-containing medium unless otherwise stated. Nominally Na+-free medium was prepared by replacing NaCl with equimolar choline chloride, and contained ∼0.04 mM Na+. BSA were excluded from media during pH measurements to promote cellular adherence to the coverslip. The pH of all media were adjusted to 7.4 using NaOH or KOH.
Oocyte and follicle handling, and microinjection
Oocytes and follicles were obtained from female CF1 mice (Charles River,St-Constant, PQ, Canada). Oocytes and follicles were isolated mechanically by fine mincing of the ovary with a razor blade, as previously described(Erdogan et al., 2005). A wave of follicular development occurs shortly after birth in mice, such that oocyte size and follicular development are related to postnatal age(Sorensen and Wassarmann,1976; Eppig,1991), allowing recovery of growing oocytes and follicles of the required size by sacrificing mice of the appropriate age. Growing oocytes and pre-antral follicles were obtained from day 10 postnatal as described previously (FitzHarris and Baltz,2006). Diameters of oocytes were determined from fluorescence images calibrated using a micrometer. At day 10, oocytes of 40-60 μm diameter were selected, and at day 20, oocytes of 70-75 μm were selected,representing the appropriate ranges for these neonates. Pre-antral follicles selected from day 10 mice had multiple layers of granulosa cells, and had not yet begun to form an antrum (secondary follicles). This population of follicles was ∼95-115 μm in diameter (oocyte plus surrounding granulosa cells). Where used, germinal vesicle-stage (GV) oocytes were obtained from primed adult female mice approximately 44 hours after equine chorionic gonadotropin injection as described previously (5 IU, intraperitoneally)(Philips et al., 2002).
Granulosa shells were prepared from pre-antral follicles of 10-day-old mice. Follicles were harvested following mincing of the ovary, as described above. Following a 30-60 minute incubation period, which rendered them less sticky, oocytes were removed from within the follicles by carefully applying pressure atop of the follicle with a glass micropipette. Only when the oocyte was observed to be expelled intact from the follicle, still enclosed in an unbroken zona pellucida, was the corresponding granulosa shell used for experiments. This could be accomplished for the majority of follicles. Though slightly more damaging to the structure of the granulosa shell than the oocytectomy method pioneered by Buccione and coworkers(Buccione et al., 1990), this method has the advantage that the oocyte remains intact when removed from the follicle, minimising contamination of the granulosa cells with ooplasm. NHE4 mRNA, which always yielded a strong amplicon in denuded oocytes by RT-PCR (see Results), was not detected in granulosa shells, indicating that contamination of the shell by oocyte cytoplasm-derived material was minimal.
Oocyte microinjection was performed using a microinjection apparatus(Harvard Apparatus-Holliston, MA, USA) and micromanipulators (Narishige)mounted on a Zeiss Axiovert microscope, as previously detailed(FitzHarris and Baltz,2006).
Measurement of intracellular pH
pHi measurements were performed using a quantitative fluorescence imaging microscopy system (Inovision, Durham, NC, USA). pH was measured using SNARF-1 which was either microinjected in dextran-coupled form(SNARF-dex; 10 kDa, estimated final concentration 0.5-1 mM), or loaded as the acetoxymethyl ester derivative (SNARF-AM 5 μM, 30 min), as previously described (FitzHarris and Baltz,2006). SNARF was illuminated using 535 nm light, and emission monitored at 600 and 640 nm. The ratio of the two intensities (640/600) was calculated after background subtraction. Where shown, exemplar images are of the 640 nm emission. Calibration was performed using the nigericin/high K+ method with valinomycin added to collapse the K+gradient (Baltz and Phillips,1999).
Ammonium pulse assay for monitoring recovery from acidosis
The ability of cells to manage acid loads was determined using the ammonium pulse assay, employed extensively in mammalian oocytes and embryos(Lane et al., 1998; Steeves et al., 2001; Harding et al., 2002). Cells were exposed to medium containing 35 mM NH4Cl for 10 minutes. This NH4Cl pulse initially causes an essentially instantaneous alkalosis as a result of rapid equilibration of NH3 across the plasma membrane. Subsequently, a more gradual pHi decrease occurs as a result of much slower equilibration of NH4+. Upon removal of NH4Cl, NH3 exits the cell, leaving behind any H+ which initially entered the cell as NH4, resulting in net intracellular acidification. In the present study, the 10-minute NH4Cl pulse was followed immediately in all cases by a 10 minute exposure to NH4Cl-free, Na+-free medium, which allows the Na+-dependence of acidosis correction to be established.
Detection of mRNAs in oocytes and granulosa cells using RT-PCR
RNA extraction was performed using an RNAeasy Micro kit (Qiagen), and reverse transcription performed using a Retroscript kit (Ambion), according to manufacturers' instructions. RNA extraction and reverse transcription were performed upon groups of no fewer than 60 oocytes or between eight and 25 granulosa shells at a time. PCR was performed using HotStarTaq PCR kit(Qiagen) on an appropriate amount of cDNA template corresponding to three oocytes, or one granulosa shell. Kidney (NHE1-4) and brain(NHE5) cDNA were used as positive controls. The amount of positive control tissue cDNA used was chosen to approximately correspond with that of three oocytes based upon measurement of RNA content of control tissue RNA preparations, and the known RNA content of one oocyte (∼0.6 ng)(Sternlicht and Schultz,1981). The same amount of positive control cDNA was used for comparison with granulosa shells, one granulosa shell having approximately the same volume as three fully grown oocytes. Samples of the final drops of medium in which oocytes were washed were subjected to the same reverse transcription protocol and used as negative controls. PCR was repeated on a minimum of two different oocyte preparations, or three different granulosa cell preparations. Amplicons were visualized on a 1% agarose gel containing 0.13 μg/ml ethidium bromide. All products were of the predicted size according to their position on the gel, and the identities of products from all five primer pairs were confirmed by direct sequencing.
Primers used for PCR were as follows (5′-3′) [accession number;predicted amplicon length]. NHE1 (also known as Scl9a1):CACCAGTGGAACTGGACCTT, AAGGTGGTCCAGGAACTGTG [NM_016981.1; 372]. NHE2(Scl9a2): ATCACGGCTGCTATTGTCGTT, GTGACCCCAGTGTCCACACACA[NM_001033289; 189]. NHE3 (Scl9a3): TCAGCTAAGCTAGGCATCAACC,ATGGTGTCAGGCGGCGGAAGT [AK033564; 412]. NHE4 (Scl9a4):CCGGAGGAACCTGCCAAAATC, TCTTCAGGAGAAAGCCGCTTGA [NM_177084; 162]. NHE5(Scl9a5): CCTCCCTGTTTGTGGTCAGT, TATGGGAGATGTTGGCTTCC [NM_138858;235]. Primers for NHE1-4 were designed based upon mouse mRNA sequences. Primers for NHE5 were generated by `blasting' the sequence for rat NHE5 mRNA (NM_138858) against the mouse genome. Primers were then designed using the matching mouse sequence. PCR products were sequenced and similarity to predicted sequences were as follows: NHE1 98%; NHE2 97%; NHE3 100%; NHE4 96%; NHE595%.
Where rates of pHi change were calculated, individual oocytes and/or granulosa shells within a given replicate were pooled to provide an average rate, and the mean±s.e.m. of the different replicates presented. Rates were calculated using SigmaPlot 8.0. Data points presented are the mean±s.e.m. of all replicates performed. Means of replicates were compared using t-tests (two groups) or ANOVA (three or more groups). Where ANOVA was used, Tukey-Kramer's and Dunnett's post-hoc tests were applied as appropriate (using Instat; GraphPad, San Diego, CA, USA).
Oocytes acquire the capacity to recover from acidosis during growth by activating Na+/H+ exchange
We first set out to determine whether denuded growing oocytes are capable of recovering from acidosis. Two separate populations were examined; fully grown oocytes (70-75 μm in diameter) obtained from 20-day-old mice, and smaller, mid-growth-phase oocytes (40-60 μm) obtained from 10-day-old mice(see Materials and methods). Oocytes were freed of their granulosa cells, and loaded with the pH-sensitive fluorophore SNARF. Acidosis was induced by exposing oocytes to a 10 minute pulse of 25 mM NH4Cl(NH4Cl pulse assay; see Materials and methods for detailed explanation). This protocol caused substantial net acidosis in oocytes of all sizes (∼0.5 pHU; Fig. 1). There was minimal recovery from the induced acidosis in the subsequent 10 minute period during which Na+ was absent from the bathing medium. Replacement of Na+ caused a rapid pHi increase in fully grown oocytes (initial rate of recovery 0.065±0.003 pHU/minute), the majority (78%) fully recovering from acidosis within 10 minutes. In sharp contrast, little or no recovery from acidosis was made by mid-growth oocytes(0.010±0.002 pHU/minute; Fig. 1). To determine whether Na+/H+ exchange may account for the Na+-dependent recovery from acidosis in fully grown oocytes, acidosis was induced in the presence of amiloride, a broad-spectrum inhibitor of Na+/H+ exchangers. Amiloride caused a substantial and highly significant inhibition of acidosis recovery (initial recovery rate 0.023±0.004 pHU/minute; P<0.002), although there was also a small amiloride-insensitive component of recovery(Fig. 1C). Oocytes thus acquire the ability to regulate against acidosis during growth by activating mechanisms which raise pHi, of which Na+/H+exchange is the dominant component.
Characterization of Na+/H+ exchange in fully grown oocytes
We next sought to determine the molecular identity of the Na+/H+ exchanger(s) which alleviate acidosis in fully grown denuded oocytes. At least nine isoforms of the Na+/H+ exchanger exist (NHE1-9), five of which (NHE1-5)are known to reside in the plasmalemma and may therefore participate in cytoplasmic pHi regulation(Orlowski and Grinstein, 2004; Nakamura et al., 2005). We first carried out RT-PCR to determine which of these isoforms are present in oocytes at the mRNA level. Primers were designed to specifically amplify cDNA generated from NHE1-5. All five primer sets generated distinct amplicons of the predicted size from positive control cDNA, and no amplicons were detected in negative controls. Amplicons of the predicted sizes were always detected for NHE1, NHE3 and NHE4, but not NHE2 or NHE5, both in oocytes from day 20 mice (fully grown oocytes) and in oocytes from day 10 mice (mid-growth phase oocytes; Fig. 2A).
We next wanted to take a functional approach to determine which NHE isoform(s) regulates against acidosis in denuded oocytes. Although specific inhibitors of HCO3-/Cl- (AE) and Na+ HCO3-/Cl- (NBC) exchangers remain elusive, a number of amiloride derivatives with isoform-selectivity for NHEs are available (Masereel et al.,2003). Since our previous experiment revealed a small amiloride-insensitive component of pHi recovery in fully grown denuded oocytes, we carried out this further characterization of Na+/H+-exchange-mediated acidosis correction in HCO3--free medium to inactivate Na+HCO3-/Cl- exchange, which may account for the amiloride-insensitive component. As in bicarbonate-containing medium, only large oocytes made a significant recovery from acidosis following NH4Cl exposure (Fig. 2B). Variation of recovery rates between oocytes within a given experiment was accentuated in bicarbonate-free medium, with a number of oocytes (31%) failing to recover from acidosis within each experiment. Nonetheless, the majority of oocytes (69%) recovered fully within 20 minutes,and the average rate of recovery was not significantly less than in bicarbonate-containing medium (0.056±0.02 pHU/minute; P>0.1). Recovery was completely inhibited by amiloride in HCO3--free conditions, consistent with the notion that the small amiloride-insensitive component seen in the presence of HCO3- might be Na+HCO3-/Cl- exchange(Fig. 2B).
Next, to identify the origin of Na+/H+ exchange activity, two isoform-selective Na+/H+-exchange inhibitors were employed. Cariporide (HOE 642) is a highly specific NHE1 inhibitor with published IC50s in the region of 0.3-3.0 μM,which has been used to identify NHE1-dependent pHi regulation in a variety of cell types (Masereel et al.,2003). S3226 exhibits a high level of selectivity for NHE3 at low concentrations (IC50s 0.02-0.7 μM), and at higher concentrations also inhibits NHE1 (IC50s ∼3 μM)(Schwark et al., 1998). Both drugs are well characterized, and their effective range in various cell types has been extensively reported (Schwark et al., 1998; Masereel et al.,2003). They have only been used previously in early mouse embryos to attempt to distinguish NHE isoforms responsible for Na+transport necessary for blastocyst formation(Kawagishi et al., 2004).
Here, cariporide potently inhibited the Na+-dependent recovery from acidosis in fully grown denuded oocytes, with substantial inhibition achieved at 1 μM, indicating that inhibition of NHE1 is sufficient to prevent acidosis recovery (Fig. 2C,D). Complete inhibition was also achieved by 10 μM S3226,which would be expected to inhibit both NHE1 and NHE3. However, 0.1 μM S3226 had no effect (Fig. 2C,D). These experiments indicate that recovery from acidosis in fully grown denuded oocytes occurs principally via NHE1.
Recovery from acidosis in granulosa cells from pre-antral follicles
Previously, it was found that small growing oocytes are only capable of recovering from an induced alkalosis when they remain within the follicle. Since granulosa cells have robust HCO3-/Cl-exchange, we proposed that gap junctions may allow the granulosa cells to regulate the pH of the ooplasm (FitzHarris and Baltz, 2006). We therefore wanted to determine whether granulosa cells also possess significant acidosis-correcting capacity which might be able to control oocyte pHi in the acid range. To address this, intact shells of granulosa cells were prepared by removing the oocyte from preantral follicles collected from 10-day-old mice(Fig. 3A; see Materials and methods for details). Granulosa shells were loaded with SNARF, and recovery from acidosis monitored using the NH4Cl-pulse approach, as for oocytes. NH4Cl treatment resulted in substantial net acidosis, as in oocytes (Fig. 3B). Na+ replacement triggered a rapid recovery from acidosis in all cases (0.064±0.011 pHU/minute), which was substantially inhibited by amiloride (initial rate of recovery 0.023 pHU/minute; P<0.05; Fig. 3B,C). Thus, granulosa cells possess robust Na+/H+ exchange which is their dominant means of correcting acidosis.
Characterization of granulosa cell Na+/H+exchange
Next, RT-PCR was carried out to establish which NHE mRNAs are expressed in granulosa shells. Distinct amplicons were produced by primers on all occasions only by NHE1 and NHE3(Fig. 4A). Cariporide and S3226 were then used to determine which NHE isoforms are responsible for the Na+-dependent, amiloride-sensitive recovery from acidosis in granulosa cells. As for oocytes, characterization of Na+/H+ exchange was carried out in HCO3--free medium to inhibit any Na+-dependent HCO3-/Cl- exchange which may be present. Somewhat unexpectedly, under these conditions there was always a slow but steady recovery from acidosis in the absence of Na+ (0.0233±0.004 pHU/minute; Fig. 4B,C). Nevertheless,pHi increased very rapidly when Na+ was replaced(0.085±0.009 pHU/minute), such that the Na+-dependent component could easily be determined (Fig. 4B). As expected, the Na+-dependent component was inhibited by amiloride, confirming Na+/H+ exchange(Fig. 4B). However, neither 1μM nor 10 μM cariporide inhibited the Na+-dependent component(Fig. 4C,D). By contrast,whereas 1 μM S3226 had no effect, recovery was substantially inhibited by 10 μM S3226 (Fig. 4C,D). Given the previously reported IC50s of S3226 for NHE1 and NHE3, we reasoned that the ability of 10 μM S3226 (but not 1 μM S3226 or 10 μM cariporide) to prevent acidosis recovery in granulosa cells may be due to simultaneous activities of NHE1 and NHE3 that are both inhibited at high S3226 concentration. To test this possibility, we performed NH4Cl-pulse experiments in the simultaneous presence of 10 μM cariporide and 1 μM S3226, treatments expected to inhibit NHE1 and NHE3, respectively, but which individually had no effect upon granulosa cells. This cariporide-S3226 co-treatment inhibited acidosis recovery to a similar degree as amiloride or 10 μM S3226 (Fig. 4C,D). Therefore, recovery from acidosis in granulosa cells is inhibited by amiloride, 10 μM S3226, or 1 μM S3226 + 10 μM cariporide, all of which are expected to block both NHE1 and NHE3 simultaneously. On the contrary, granulosa cells remain capable of recovering from acidosis if only one isoform is inhibited.
Characterization of Na+-independent recovery from acidosis in granulosa cells
We next wanted to identify the Na+-independent means of acidosis recovery which was evident in granulosa cells in HCO3-free medium. This mechanism was uninhibited by any of the NHE inhibitors (see Fig. 4B,C). In experiments in which Na+ was not returned to the medium, granulosa cell pH recovered almost fully within 20 minutes of NH4Cl removal(78±4% recovery; Fig. 5A). One possibility was that this slow pHi increase may be attributable to a V-type H+-ATPase, which are commonly found on intracellular membranes, but have been found on plasmalemmae and shown to correct acidosis in Na+- and HCO3--free conditions in some cells (Merzendorfer et al., 1997; Kawasaki-Nishi et al., 2003). N-ethylmaleimide (NEM; 100 μM), a widely employed nonspecific inhibitor of V-ATPases, abolished the Na+-independent recovery from acidosis (initial rate of recovery 0.001±0.003 pHU/minute compared with controls 0.023±0.004 pHU/minute; P<0.01; Fig. 5A). However, subsequent re-addition of Na+ did not trigger a pHi increase in these experiments, implying that Na+/H+ exchange was also incapacitated by NEM(Fig. 5A). Though Na+/H+ exchange is clearly not responsible for the Na+-independent pH increase, this result raises questions about the specificity of NEM in these experiments. We therefore also employed bafilomycin and concanamycin, two related macrolide antibiotics that are highly specific for V-ATPases (Bowman et al., 1988; Drose and Altendorf, 1997). Inhibition of proton transport by nanomolar concentrations of concanamycin or bafilomycin is a strong indication of a role for V-ATPase (Drose and Altendorf,1997). Bafilomycin dramatically inhibited the Na+-independent recovery (Fig. 5A,B; rate of recovery 0.005±0.001 pHU/minute; P<0.01), similar to NEM. Concanamycin caused a partial inhibition(rate 0.013±0.001 pHU/minute). Importantly, re-addition of Na+ triggered a robust pHi increase in the presence of either antibiotic (Fig. 5A). We therefore conclude that the Na+- and HCO3--independent pHi increases observed in granulosa cells following induction of acidosis are attributable to a H+-transporting V-ATPase.
Gap junctions may provide follicle-enclosed oocytes with access to mechanisms that correct acidosis
Having characterized the mechanisms that drive acidosis recovery in granulosa cells, it was now important to determine whether the small growing oocyte, which by itself is incapable of recovering from acidosis, can access these mechanisms. To do this, pHi was measured in small (40-60μm diameter) denuded and follicle-enclosed oocytes by microinjecting a cohort of each with SNARF-dextran (10 kDa). Using this approach, the pHi of follicle-enclosed oocytes can be monitored in the absence of contaminating fluorescence from the follicle cells, allowing simultaneous comparison with denuded oocytes (see Fig. 6). The initial `resting' pH of follicle-enclosed oocytes(7.26±0.03) was significantly higher than that of denuded counterparts(7.02±0.02; P<0.01). Little or no recovery from acidosis occurred following NH4Cl treatment in denuded oocytes, as in previous experiments (0.005±0.001 pHU/minute). However, pHiincreased substantially in follicle-enclosed oocytes upon re-addition of Na+ (initial rate 0.038 ±0.005 pHU/minute), indicating that the granulosa cells afford the oocyte the capacity to recover from acidosis(Fig. 6A).
To establish whether gap junctions are involved in this,18α-glycyrrhetinic acid (AGA), a small-molecule gap junction inhibitor,was employed. We have previously shown that a 10-minute incubation with AGA completely inhibits oocyte-granulosa cytoplasmic continuity as determined by transfer of a fluorescent dye (FitzHarris and Baltz, 2006). Here, AGA significantly inhibited the Na+-dependent acidosis recovery (identified above as due to Na+/H+ exchanger activity) in follicle-enclosed oocytes compared to controls (P<0.01; Fig. 6B,C). Moreover,pHi increased steadily during the Na+-free period in control (vehicle) follicle-enclosed oocytes (0.0189±0.001 pHU/minute),and this Na+-independent pH increase (identified above as due to V-ATPase activity) was also significantly inhibited by AGA(0.0074±0.001 pHU/minute; P<0.01; Fig. 6C). AGA did not inhibit acidosis recovery in granulosa shells, or denuded fully grown germinal vesicle-stage oocytes, indicating that the ability of AGA to block acidosis recovery in follicle-enclosed growing oocytes is unlikely to be due to a unexpected effect on pHi regulation(Fig. 6B, right). Thus, whereas small denuded oocytes are unable to recover from acidosis, follicle-enclosed oocytes possess acidosis-correcting activities that resemble those in the granulosa cells, provided gap junctions remain functional.
We also used octanol, another compound commonly employed as a gap junction inhibitor. Like AGA, octanol also eliminated the ability of follicle-enclosed oocytes to recover from acidosis, eliminating both the Na+-dependent and Na+-independent components. However,we found that octanol also inhibited the Na+-dependent component of acidosis recovery in isolated granulosa shells to a varying extent, although the Na+-independent component was unaffected. Therefore, whereas the experiments with octanol (not shown) were consistent with the conclusion that the enclosed oocyte pHi is regulated by the surrounding granulosa cells, since the Na+-independent (V-ATPase-dependent)recovery was not affected by octanol in granulosa shells but was eliminated in the enclosed oocyte, the apparent direct inhibition of Na+/H+ exchange by octanol in granulosa cells precluded its use in examining the Na+-dependent recovery.
Simultaneous measurement of pHi in oocytes and granulosa cells
The most reasonable interpretation of the experiments presented hitherto is that gap junctions allow oocytes access to acidosis-correcting mechanisms within the granulosa cells. Within such a model we would predict that, in intact oocyte-granulosa complexes, recovery from acidosis should take place at least as quickly in granulosa cells as in the oocyte. To examine this directly, it was necessary to simultaneously monitor recovery from acidosis in oocytes and their surrounding granulosa cells, which we have done previously using partly denuded oocytes (FitzHarris and Baltz, 2006). Since it was not technically possible to obtain partly denuded oocytes from growing follicles, we instead used fully grown GV oocytes from PMSG-primed adult mice. Part-enclosed oocytes and their accompanying cumulus granulosa cells adopted a markedly higher pHithan denuded oocytes, consistent with previous data(FitzHarris and Baltz, 2006). Recovery from acidosis occurred more rapidly in partly-enclosed oocytes. Strikingly, the Na+-independent phase of recovery from acidosis,which we characterized in granulosa cells and follicle-enclosed oocytes, was evident in oocytes and accompanying granulosa cells in these experiments such that oocytes and accompanying granulosa recovered synchronously(0.018±0.005 pHU/minute), whereas pHi increase in Na+-free medium in denuded oocytes imaged simultaneously was minimal (0.007±0.003 pHU/minute; P<0.05; Fig. 7). In addition,pHi recovery following Na+ replacement occurred more rapidly in partly enclosed oocytes in each experiment. These data are consistent with the other experiments presented here, indicating that oocytes have access to the acidosis-correcting mechanisms of the granulosa cells. In addition, this confirms that both the Na+-dependent and -independent phases of recovery continue to be present in granulosa cells even when the oocyte is fully grown.
The major finding here is that growing oocytes require granulosa cells and functional granulosa-oocyte gap junctions to correct acidosis. Fully grown denuded oocytes recover from an induced acidosis, but smaller oocytes lack this ability. When taken together with the previous finding that denuded small growing oocytes are incapable of correcting alkaline loads(Erdogan et al., 2005), this establishes the growing oocyte as a rare example of a mammalian cell unable to regulate its own pHi. However, in the more physiological situation where the small oocyte is coupled to its granulosa cells, the pHiof the oocyte can be maintained. The finding that small oocytes are only capable of raising their pHi when coupled to granulosa cells explains the finding that small oocytes adopt a lower `resting' pH when denuded (FitzHarris and Baltz,2006) (Fig. 6). The low pHi probably reflects the effect of electrochemical equilibrium or metabolic H+ generation on oocyte pHi in the absence of pHi-regulatory mechanisms and, as such, should be considered an artifact of removal from the ovary. As the oocyte grows it acquires the ability to regulate against acid loads, such that fully grown oocytes are capable of independently maintaining the appropriate pHi when oocyte-granulosa gap junction communication is broken after ovulation is triggered.
Distinct pH regulatory mechanisms in oocytes and granulosa cells
Using isoform-specific inhibitors, we were able to establish that oocytes acquire the ability to regulate against acidosis independently of granulosa cells by developing NHE1-mediated Na+/H+ exchange activity. An examination of pHi regulatory mechanisms was also carried out on granulosa cells from 10-day-old mice, at which stage the oocyte relies upon the granulosa cells for pHi regulation. Granulosa cells were found to possess several distinct mechanisms. Robust Na+/H+ exchange was only inhibited by treatments that simultaneously inactivated NHE1 and NHE3. These data imply that either isoform on its own is capable of driving maximal recovery from induced acidosis,indicating a functional redundancy between the two isoforms in granulosa cells. Two populations of distinctly regulated antiporters may also allow precise pHi regulation during folliculogenesis. Consistent with the idea that NHE1 and NHE3 together account for Na+-dependent,HCO3--independent pHi regulatory capacity in granulosa cells, NHE1 and NHE3 mRNAs were the only isoforms detected in granulosa cells by RT-PCR. The finding that a small proportion of Na+-dependent recovery was amiloride-insensitive in HCO3--containing, but not HCO3--free, medium, suggests that granulosa cells may also possess a minor component of Na+,HCO3-/Cl- exchange, however, we have not investigated this further. An additional component of acidosis recovery was present in granulosa cells, which was HCO3- and Na+ independent, uninhibited by Na+/H+exchange inhibitors, and inhibited by bafilomycin, concanamycin and NEM,indicative of a V-type H+ ATPase. It is difficult to predict to what degree the V-ATPase contributes to pHi homeostasis in vivo, as this component was less evident in HCO3--containing medium, most likely masked by HCO3-/CO2-based cytoplasmic buffering. Nonetheless, murine granulosa cells should be added to the relatively limited list of cells in which V-type H+ ATPases have been shown capable of regulating cytoplasmic pHi(Nelson and Harvey, 1999; Merzendorfer et al., 1997; Kawasaki-Nishi et al., 2003). Avian granulosa cells possess an unidentified Na+- and HCO3--independent means of raising pHi(Li et al., 1992), suggesting that plasmalemmal V-ATPase-mediated pHi regulation may be conserved in granulosa cells. Overall, it is clear that granulosa cells have been amply equipped with mechanisms that regulate against pHi changes,reflecting their unique role as the regulators of ooplasmic pH.
Growing oocytes have access to granulosa cell pHiregulation
The current data strongly support a model in which gap junctions allow pHi-regulatory mechanisms within granulosa cells to regulate oocyte pHi. Strikingly, both the Na+-dependent (NHE) and Na+-independent (V-ATPase) phases of acidosis recovery evident in granulosa shells were echoed by granulosa-enclosed oocytes, and both were abrogated by gap-junction inhibition. The appearance of the Na+-independent phase in follicle-enclosed oocytes is compelling,as we found no evidence that the oocyte itself ever exhibits pHi-regulatory V-ATPase activity. Unfortunately, the presence of NHE1 in granulosa cells and denuded oocytes prevented us from selectively inhibiting granulosa cell Na+/H+ exchange in intact follicles without affecting the Na+/H+ exchange present in fully grown oocytes, an experiment which might have added evidence for the direct action of granulosa cell exchangers upon ooplasm pHi. Nevertheless, we have been unable to prevent recovery from acidosis in follicle-enclosed oocytes using even 10 μM cariporide (data not shown), a result which would not be expected if the function of the granulosa cells were to activate oocyte transporters, since Na+/H+ exchange in fully grown oocytes was potently inhibited by 1 μM cariporide (see Fig. 2). Taken together, the evidence suggests that gap junctions allow the granulosa cells to serve as an extension of the oolemma, regulating ooplasmic pHi on its behalf. Therefore the ovarian follicle provides the first demonstration that gap junctions can allow one cell to regulate the pHi of its neighbour in a physiological setting, a notion that is summarised in Fig. 8. Though the identity of the molecules that traverse the gap junctions allowing pHi to equilibrate between the oocyte and the granulosa cells is unknown, direct transfer of free protons is unlikely given the high buffering capacity of the cytoplasm (Swietach et al.,2003). Based on the rate of transfer of injected acid between gap junction-coupled cardiomyocytes it has been proposed that the mobile buffer should be around 200 Da, in which case taurine, inorganic phosphate, or dipeptides could be candidates (Zaniboni et al., 2003), and bicarbonate could also contribute under physiological conditions.
pHi regulation during oocyte growth
To conclude, the growing mammalian oocyte is incapable of regulating its own pHi, and has `outsourced' this function to the granulosa cells,though the physiological benefit of this arrangement remains to be determined. Although Na+/H+ and HCO3-/Cl- exchange are passive secondary transporters, ATP hydrolysis is indirectly required to establish necessary transplasmalemmal ionic gradients. It may be that, during oocyte growth,energy resources are instead reserved for the metabolically expensive process of cell growth, or that reduced demands upon oocyte metabolism minimises potentially damaging effects of free radical production during the prolonged(tens of years in humans) meiotic arrest prior to recruitment into the growth phase.
This work was funded by a Canadian Institutes of Health Research Operating Grant (MOP74515) to J.M.B., and a CIHR Fellowship to G.F. The authors wish to thank Guillaume Halet and John Carroll for critical reading of the manuscript. We thank Dr Jürgen Pünter (Sanofi Aventis, Frankfurt) for the kind gifts of cariporide and S3226.