ABSTRACT
The ion physiology of osmotic swelling and a consequent loss of epithelial patency was examined in the ovaries of the moth Hyalophora cecropia. After 30 min in the presence of an activator of cyclic-AMP-dependent protein kinase (PKA), the membrane potentials of both oocyte and follicle cells had hyperpolarized by approximately 30 %, cytoplasmic pH had dropped from 7.26 to 7.06, a normally low Cl− conductance had increased and the follicle cells had begun to swell. Since ion distribution studies have indicated that conductance increases should depolarize membranes in this system, it is proposed that hyperpolarization may be effected by an azide-inhibitable component of the membrane potential. Nanomolar levels of bafilomycin, an inhibitor of H+ V-ATPase, blocked the active component and prevented osmotic swelling in response to PKA activation. Under a variety of circumstances, correlations were seen between membrane potential and cytoplasmic pH, suggesting that substrate availability to the proton pump may contribute to hyperpolarization. H+ V-ATPases are known to energize ion and water transport across many epithelia, but in this case they generate water absorption by the epithelium. The increase in Cl− conductance was also required for the swelling response: the Cl− channel blocker anthracene-9-carboxylic acid prevented both swelling and hyperpolarization, as did Cl− substitution in the medium. Differences in isotope loading rates between 36Cl− and 86Rb+ suggested that, after PKA activation, Cl− functions other than as a counterion for K+ uptake.
Introduction
In ovarian follicles of the cecropia moth, vitellogenesis ends when the follicle cells surrounding the oocyte transform from a porous layer into a more conventional tight epithelium. The transformation entails three changes that prevent further access of hemolymph proteins to the oocyte surface. The follicle cells dismantle a sulfated glycosaminoglycan matrix that contributes to patency during vitellogenesis (Telfer, 1979), they close the intercellular spaces by increasing their volume, and they form tight-junction-like diffusion barriers at their apical ends (Rubenstein, 1979).
Cell-permeant analogs of cyclic AMP stimulate follicle cells to undergo these changes in vitro by activating cyclic-AMP-dependent protein kinase (PKA) (Wang and Telfer, 1996, 1997). After a lag period of over an hour, vitellogenin uptake stops, the epithelium becomes impermeable to small molecules as well as to vitellogenin, and production of the intercellular matrix is inhibited. Most relevant to the present study, cellular swelling also occurs, so that each cell now presses against its lateral neighbors.
In a number of other systems, transepithelial water and ion transport are energized by a proton-translocating, vacuolar-type ATPase (H+ V-ATPase) (Harvey and Wieczorek, 1997; Harvey et al. 1998) that is selectively inhibited by nanomolar concentrations of bafilomycin (Bowman et al. 1988; Drose and Altendorf, 1997). We present evidence here that the hydration phase of development in Hyalophora cecropia follicles, an example of developmentally regulated cellular swelling, also requires an H+ V-ATPase. In particular, we propose that membrane hyperpolarization, an obligatory early step towards epithelial swelling, results from an increased concentration of protons available to a bafilomycin-sensitive mechanism. An early increase in membrane conductance, which could be inhibited with 9-AC and depended on the presence of Cl− in the medium, was also an obligatory step in the pathway towards swelling.
Materials and methods
Preparation and incubation of follicles
Vitellogenic follicles were dissected from Hyalophora cecropia (L.) females showing early wing pigmentation (day 18 of the 22 day pupal–adult molt at 25 °C). Follicles were dissected in a 300 mosmol l−1 physiological salt solution (PSS) empirically developed to optimize yolk formation and electrophysiological stability (Anderson and Telfer, 1969; Woodruff et al. 1992). It contained 40 mmol l−1 KCl, 15 mmol l−1 MgCl2, 4 mmol l−1 CaCl2, 110 mmol l−1 Tris-succinate (pH 6.5) and 70 mmol l−1 sucrose. The composition of PSS reflects the unusual inorganic ion composition of the hemolymph in Lepidoptera whose larvae develop on Na+-poor, K+-rich diets (Harvey et al. 1975). Chloride substitution experiments utilized the gluconate salt of K+ and the SO42− salts of Mg2+ and Ca2+. Choline chloride was used in place of KCl in K+ substitution experiments.
For each experiment, matched chains of ten vitellogenic follicles were dissected from the eight ovarioles of a single female. Hemolymph proteins were soaked out of the intercellular spaces by two 20 min rinses with shaking in PSS. For PKA activation, the Sp-isomer of adenosine-3′,5′-monophosphorothioate (Sp-cAMPS) was used at 1 mmol l−1. This cyclic AMP analog is cell-permeant, resistant to degradation by phosphodiesterases and promotes dissociation of the regulatory and catalytic subunits of PKA (Rothermel and Botelto, 1988). At the concentrations used, Sp-cAMPS induces the termination of vitellogenin uptake with a lag period of approximately 1 h (Wang and Telfer, 1996).
Each chain of ten follicles was incubated in 200 μl of PSS containing specified isotopes, Sp-cAMPS and/or inhibitors. Stock solutions of inhibitors with low solubility in water (bafilomycin and anthracene-9-carboxylic acid, 9-AC) were prepared in dimethylsulfoxide (DMSO) at concentrations high enough so that the final incubation medium contained less than 1 % DMSO. The same concentration of DMSO was added to control cultures not exposed to these inhibitors. For measurements of electrical properties and proton or isotope contents, the chain of ten was divided into individual follicles and each result is expressed as a mean of the ten (± S.E.M.).
Isotope loading and unloading
Isotopic ion loading was measured by incubating follicles in 36Cl− (3.7 MBq ml−1) or 86Rb+(3.7 MBq ml−1). After incubation, the chain was dropped into 50 ml of PSS for several seconds, during which individual follicles were separated and transferred for counting to scintillation vials containing 250 μl of a tissue solubilizer (50 mmol l−1 Tris-HCl, pH 7.0, 1 % SDS and 0.35 % 2-mercaptoethanol). To measure unloading rates, follicles were first incubated in 36Cl− for 2.5 h, by which time the amount of label per follicle had reached a steady state. They were then transferred to 5 ml of fresh PSS without isotope. After 10 min, the follicles were individually dissolved and counted as in the loading experiments.
Electrophysiology
Electrical measurements were made in a 0.25 ml open-topped chamber grounded through a 3 mol l−1 KCl bridge to a Ag/AgCl wire (Woodruff et al. 1992). Standard 3 mol l−1 KCl microelectrodes were attached to S-7071A electrometers (World Precision Instruments, New Haven, CT, USA) for membrane potential measurements or to an S-7071A ionophoresis unit (also from World Precision Instruments) to provide current pulses. Proton-selective microelectrodes were constructed using a liquid ion-exchange resin from World Precision Instruments, as described by Ammann et al. (1981) and Palmer and Civan (1977). Calibrations and measurements were performed as in Woodruff et al. (1992). Prior to impalement with microelectrodes, follicles were treated for 1 min with 1 mg ml−1 collagenase (type 1A, Sigma) to soften the basement membrane (Woodruff and Telfer, 1990).
At the stages studied, the cytoplasmic bridges to the nurse cells have disintegrated, and an epithelium of follicle cells completely surrounds the oocyte. The follicle cells thus form a closed ellipsoid, differing in this regard from the tubular or flat-sheet configurations of the more intensively studied insect epithelia. A second important difference is that, prior to treatment with Sp-cAMPS and during the subsequent lag phase, oocyte and follicle cells communicate via ion- and fluorescent-dye-permeable gap junctions (Woodruff, 1979). Properties measured with microelectrodes or isotopes in one type of cell therefore reflect the contributions of both.
Follicle cell swelling
Epithelial cell height was used as the index of swelling (Wang and Telfer, 1997). It was measured in living follicles viewed through the low-power objective lens of an inverted microscope with the focal plane set halfway through the follicle, where its diameter appeared widest. Epithelial height was measured with a filar micrometer as the distance at this level between the edge of the opaque yolk mass and the basement membrane forming the surface of the follicle.
Reagents
The Sp-isomer of adenosine-3′,5′-monophosphorothioate (Sp-cAMPS) was purchased from Biolog Life Science Institute (Bremen, Germany); 36Cl− and 86Rb+ were purchased from New England Nuclear (Newark, DE, USA); bafilomycin-A1 was the gift of Dr Philip Rea, University of Pennsylvania; anthracene-9-carboxylic acid (9-AC) and all other reagents came from Sigma (St Louis, MO, USA).
Results
Early responses to Sp-cAMPS
Three changes were detected during the lag phase in Sp-cAMPS-induced termination of vitellogenin uptake. In an earlier report, the height of the follicle cells at the end of this 1.5 h period was, on average, 20 % greater than that of follicles incubated in PSS alone, and cellular cross section was 25 % greater, yielding a calculated volume increase of 50 % (Wang and Telfer, 1997). In the time course study shown in Fig. 1, the increase in height was first detected after 30 min in Sp-cAMPS and reached a stable, maximum value by 3 h. Epithelial height in treated follicles was 20 % greater than that in follicles incubated for 2 h in PSS without Sp-cAMPS. Cell widths were not measured but, as in the earlier study, swelling had obliterated the intercellular spaces by this time, so that a volume increase of 50 % can be assumed here also. Since the overall dimensions of the follicles studied did not change detectably, epithelial swelling must have occurred primarily at the expense of the intercellular spaces.
A plot of current injected versus voltage change indicated that, within 30 min of treatment with Sp-cAMPS, there was an increase in general membrane conductance (filled squares versus filled diamonds in Fig. 2). The difference largely disappeared in Cl−-free PSS (open circles), but was unaffected by K+-free PSS (open triangles). As would be expected, the increase in Cl− conductance was accompanied by increases in the loading and unloading rates of 36Cl−. The amount of isotope taken up in 10 min began to increase within 15 min of exposure to Sp-cAMPS and continued to rise over the next 1 h (Fig. 3A). Unloading accelerated on a similar time course (Fig. 3B). The Cl− channel blocker 9-AC (Palade and Barchi, 1977; Welsh, 1984) had no effect on 36Cl− uptake in unstimulated follicles, but prevented the increase that occurred in response to Sp-cAMPS (Table 1).
As might be expected for fluid uptake, the increase in follicle cell height required Cl− and K+, the major inorganic ions in PSS. In this experiment, Sp-cAMPS elicited a 25 % greater follicle cell height during incubation for 2 h in PSS, whereas in Cl−-free PSS, swelling was reduced to 6 % and in K+-free PSS it was reduced to 3 % (Table 2).
The dependencies of swelling on Cl− and K+ were based on separate principles, however, since Sp-cAMPS affected labeling with 36Cl− and 86Rb+ very differently. For the first hour of incubation in 86Rb+, Sp-cAMPS did not significantly affect the amount of label accumulated (Fig. 4A), but uptake subsequently accelerated, as would be expected for cells that were utilizing the osmotic effects of K+ to absorb water. In contrast, 36Cl− reached the same level during 3 h in the presence of Sp-cAMPS as in its absence (Fig. 4B). Other ions must therefore neutralize the electrical effects of the increased accumulation of K+.
The early responses to Sp-cAMPS also included a temporary membrane hyperpolarization. The response remained strong until 1 h and then disappeared by 2 h. The example shown in Fig. 5 was seen with electrodes positioned in an oocyte, but the same response was seen in follicle cells (not shown). That the two compartments appeared to be coupled in this regard is not surprising since, during vitellogenesis, fluorescent compounds and ion currents readily pass between them via gap junctions (Woodruff, 1979; Woodruff and Telfer, 1990).
The basis of hyperpolarization
ΔΨ, the membrane potential measured between the cytoplasm of either the oocyte or follicle cells and PSS, includes a passive component that varies between different females from −12 to −20 mV and an azide-inhibitable, active component of −20 to −25 mV (Woodruff et al. 1992). Hyperpolarization is unlikely to result from the increase in Cl− conductance, because this would be expected to depolarize follicle membranes; in a set of vitellogenic follicles in which ΔΨ averaged −35 mV, the equilibrium potential for Cl− was −22 mV (Woodruff et al. 1992). In addition, since K+ substitution did not affect the change in conductance induced by Sp-cAMPS (Fig. 2), opening channels accessible to this ion should not account for hyperpolarization either. A change in the active component of the membrane potential is therefore more likely to underlie this response.
Bafilomycin, an inhibitor of H+ V-ATPases that acts at low nanomolar concentrations (Bowman et al. 1988; Drose and Altendorf, 1997), depolarized oocyte and follicle cell membranes to nearly the same extent as azide (Table 3). At a bafilomycin concentration of 150 nmol l−1, depolarization began in 5 min and was complete at 20 min (not shown). Inhibition of the active component was approximately 50 % in 30 nmol l−1 bafilomycin and reached a maximum at 1 μmol l−1 (Fig. 6). This result suggests that H+ V-ATPase activity becomes a candidate for energization of cell membranes in H. cecropia follicles, as has been established for goblet cells in the midgut of Manduca sexta (reviewed by Harvey and Wieczorek, 1997).
Two models of how an increase in bafilomycin-sensitive proton pumping could be elicited by Sp-cAMPS were considered. One called for a molecular modification of the pump itself or its translocation to the cell surface, and the second for an increase in the availability of protons serving as substrate. Since the two would entail opposite effects on cytoplasmic pH, an experimental distinction between them was possible: molecular modification that increased the efficiency of the pump, or insertion of new units into the cell surface, should produce a rise in cytoplasmic pH, whereas an increase in the amount of protons should be detected as a fall in pH. After 30 min in Sp-cAMPS, when hyperpolarization was at a maximum, cytoplasmic pH in the oocyte had dropped from 7.26±0.01 to 7.06±0.01 (Table 3) (P<0.001), suggesting that cytoplasmic acidification promotes hyperpolarization.
Sp-cAMPS induced partial but significant membrane repolarization in follicles treated with either 150 nmol l−1 bafilomycin or 10 mmol l−1 azide, and here again cytoplasmic acidification occurred (Table 3). Since the concentration of bafilomycin used in this experiment would not have completely suppressed the active component of ΔΨ (Fig. 6), acidification would presumably allow residual electrogenic enzyme activity to produce the observed partial repolarization. An increase in proton availability should also have enhanced electrogenic enzyme activity in 10 mmol l−1 azide, since this reagent should inhibit only mitochondrial ATP production and not the generation of ATP by glycolysis. A fundamental question arising from this model, therefore, is the origin of the protons produced in response to Sp-cAMPS.
Further effects of 9-AC
That the increase in Cl− conductance plays a pivotal role in the response to PKA activation was indicated by additional effects of 9-AC. This Cl− channel blocker prevented not only the increase in 10 min loading of Cl− (Table 1), but also hyperpolarization (Table 1, right-hand column), follicle cell swelling (Table 4) and the termination of vitellogenin uptake (not shown). Like the differing effects of PKA activation on 36Cl− and 86Rb+ uptake (Fig. 4), the variety of responses blocked by 9-AC implies that the role of Cl− in water absorption is more complex than simply as a counterion for K+ uptake.
Contrast with some water-secreting epithelia
In a number of insect epithelia concerned with fluid secretion, an amiloride-sensitive cation/H+ antiporter couples the export of cations to the down-gradient return of H+ to the cytoplasm. The secreted cation is replaced in the cytoplasm from the basal extracellular environment through any of a variety of membrane channels or pumps (reviewed in Harvey et al. 1998.) In Malpighian tubules of Rhodnius prolixus, for instance, 5-hydroxytryptamine-stimulated fluid secretion is retarded both by amiloride acting at the apical surface of the epithelium (Maddrell and O’Donnell, 1992) and by furosemide, an inhibitor of Na+/K+/Cl− cotransport at the basal surface (O’Donnell and Maddrell, 1984). However, neither amiloride nor furosemide affected Sp-cAMPS-induced fluid uptake in follicle cells (Table 4). Despite their mutual reliance on an H+ V-ATPase for energization of fluid movements, ancillary mechanisms of cation transport are presumably very different between these cases of fluid uptake and transcellular secretion.
Discussion
Sites of PKA activation and hyperpolarization
The restriction of swelling to the follicle cells during the 4 h incubations used in this study is consistent with the normal timetable of post-vitellogenic development. In situ, the oocyte also undergoes a 50 % volume increase, but not until several hours after follicle cell swelling has been completed (Y. Wang and W. Telfer, unpublished observation). That the follicle cells initiate the response to PKA activation in vitro is consistent with the finding that they contain 90 % of the total follicular PKA activity (Wang, 1996).
When sections of vitellogenic follicles from Manduca sexta were stained with monoclonal antibodies against the 20 kDa subunit of an H+ V-ATPase (Janssen et al. 1995), the stain was localized both in the cortex of the oocyte and at the outer or basal surface of the follicular epithelium. The former would be expected, since vesicular proton loading, a general feature of endocytosis, is required for the transfer of vitellogenin from endosomes to yolk bodies (Stynen et al. 1988). The follicle cell surface, by contrast, is much less active in endocytosis, and localization here is more consistent with a role for H+ V-ATPase in generating follicular membrane potentials.
Follicle cells and oocyte communicate through gap junctions (Woodruff, 1979), and any ionic changes induced in one should therefore be transmitted to the other. This supposition is supported by the observations recorded in Tables 2 and 3 that ΔΨ measured with electrodes in the follicle cells was consistently within 1–4 mV of that recorded in the oocyte. The epithelium was always the more negative of the two compartments, however, and this in turn is consistent with a role in generating the active component of ΔΨ.
The consequences of PKA activation
Vitellogenic follicles responded to incubation in Sp-cAMPS by increasing their Cl−-based membrane conductance, by cytoplasmic acidification, by a membrane hyperpolarization strong enough to overcome the expected depolarizing effects of Cl− conductance and by an increase in the rate of K+ uptake, as indicated by accelerated 86Rb+ loading. The challenge now is to determine how these changes are related to each other and to the 50 % increase in follicle cell volume.
The increase in Cl− conductance in response to cyclic AMP analogs is a fundamental change in the ion physiology of vitellogenic follicles, since ion substitution experiments have shown that Cl− contributes little to ΔΨ during yolk deposition (Woodruff et al. 1992 for H. cecropia; O’Donnell and Sharda, 1994, for Rhodnius prolixus). It is a necessary step in the response to Sp-cAMPS because, as noted above, 9-AC and Cl−-free medium blocked both hyperpolarization and follicle cell swelling.
An earlier study using ion-selective microelectrodes (Woodruff et al. 1992) showed that an increase in the conductance for Cl−, H+, Ca2+ or Mg2+ should depolarize vitellogenic follicles. The equilibrium potential for K+, the principal permeant cation, is close enough to ΔΨ so that a change in conductance would in this case have little effect. K+ substitution, in any event, indicated that this cation contributes less than Cl− to the early increase in conductance (Fig. 2). Hyperpolarization is therefore more probably due to the active fraction of ΔΨ that can be inhibited with azide.
The finding that bafilomycin depolarized cell membranes in the follicle to approximately the same degree as azide suggests that the electrogenic effects of H+ extrusion account for most of this fraction of ΔΨ. Observations on cytoplasmic pH have supported this idea. In an earlier study (Stynen et al. 1988), ΔΨ became more negative in synchrony with a fall in cytoplasmic pH when follicles were incubated with the cation/H+ exchange ionophores nigericin and monensin. The response was inhibited by azide and thus depended on metabolic energy. Interpretation of this relationship focused at the time on the ability of proton ionophores to inhibit endosome processing during vitellogenin uptake. Subsequently, O’Donnell and Sharda (1994), working with follicles of R. prolixus, showed that experimental acidification was accompanied by hyperpolarization of follicle membranes. Their analysis went on to provide evidence for a proton pump that generated approximately 10 % of the membrane potential. The fraction of ΔΨ attributable to the proton pumps in H. cecropia was greater than in R. prolixus, approximately 60 % as opposed to 10 %, but this may be a manifestation of evolutionary adaptation to a Na+-poor diet which, in moths, has led to a high-[K+]/low-[Na+] hemolymph and to the use of proton pumps to energize cell membranes in place of the more conventional ouabain-inhibitable Na+/K+-ATPase (Wieczorek et al. 1986).
The drop in pH that occurs after PKA activation should by itself promote electrogenesis since it indicates an increase in substrate availability to the H+ V-ATPase. An additional mechanism is also possible, however. In mammalian nephrons and a number of other epithelia, acidification promotes the insertion of cytoplasmic vesicles containing H+ V-ATPase into apical cell membranes of mitochondria-rich cells. In the nephron, this mechanism promotes proton removal from the blood and secretion into the urine (reviewed by Brown and Breton, 1996). While acidification following PKA activation of follicles is generated by the cells themselves, rather than being imposed externally, vesicle insertion into plasma membranes could well be an additional factor in hyperpolarization.
The steps leading to fluid uptake in response to Sp-cAMPS are summarized in Table 5. Activation of PKA by this and other analogs of cyclic AMP was demonstrated in the references indicated and was highlighted by the inability of the Rp-stereoisomer of cAMPS to block vitellogenin uptake or to induce follicle cell swelling. The increase in Cl− conductance is presumably the consequence of a phosphorylation cascade initiated by PKA activation in the follicle cells. For technical reasons, cytoplasmic acidification was measured only in the oocyte; we presume that it reflects primarily proton diffusion across gap junctions, possibly from even more strongly acidified follicle cells. For the reasons described above, hyperpolarization is presumed to result from an increased substrate availability to the electrogenic proton pump, although the possibility that the lower pH promotes vesicle insertion into the plasma membranes is not ruled out. Hyperpolarization could, in principle, account for the accelerated uptake of K+, but the specific channels involved are undefined and the possibility of another mode of energization is not ruled out. Swelling occurs first in the follicle cells because PKA and plasma-membrane-associated H+ V-ATPase are concentrated there, and perhaps also because these cells are surrounded by spaces into which they can readily expand; oocyte swelling, in contrast, would be constrained by the vitelline envelope, whose mechanical loosening would be essential for the final swelling of the follicle that is observed in situ.
Several basic questions are left unanswered by these considerations. Are the increases in Cl− conductance and acidification separate effects of PKA activation, or does one generate the other? Does acidification result from metabolic changes that increase organic acid production or from proton leakage across membranes? How is K+ uptake accelerated and how are its electrical effects neutralized? And, central to understanding hydration in the context of ovarian development in situ, how are individual follicles stimulated to generate these changes at an appropriate stage of egg formation?
ACKNOWLEDGEMENTS
This work was supported by a grant from the National Institutes of Health (GM 32909). We would also like to acknowledge the advice of William R. Harvey and Richard I. Woodruff.