ABSTRACT
In S. bullata, the ovaries contribute to the synthesis of yolk polypeptides. A specific antiserum for yolk polypeptides was used to visualize the presence of yolk polypeptides in the follicle cells during their differentiation. After vitellogenesis has started, all follicle cells contain yolk polypeptides. The squamous follicle cells covering the nurse cells and the border cells lose yolk polypeptides before mid-vitellogenesis, whereas the follicle cells over the oocyte contain yolk polypeptides until after late vitellogenesis. All follicle cells are immunonegative afterwards. In vitro translation of poly(A)+ RNA demonstrated that the presence of yolk polypeptide mRNA correlates well with follicle cell immunopositivity for yolk polypeptides. This suggests that the follicle cells synthesize the ovarian yolk polypeptides. Differences in cellular and nuclear morphology, total and poly(A)+ RNA synthesis and the rate of yolk polypeptide synthesis were shown to be correlated with the presence or absence of yolk polypeptides in the differentiating follicular epithelium. The possible relationship between these different aspects of follicle cell differentiation, follicle cell polyploidy and the extracellular current pattern around follicles are discussed.
Introduction
Meroistic polytrophic egg follicles of Sarcophaga bullata consist of 16 sibling daughter cells, surrounded by a follicular epithelium of mesodermal origin. Four follicle cell types with distinct morphology and distribution are distinguished: squamous, columnar, border and polar cells (Geysen & De Loof, 1983). Although polarized morphological differentiation of the follicular epithelium is a common feature of polytrophic follicles (King & Aggerwal, 1965; Cummings & King, 1969; Chia & Morrison, 1972; Khurad & Thakabe, 1980), the idea that they probably deal with different functions has not been elaborated until recently. It was suggested that the follicle cells covering the nurse cells are involved in generating extracellular current patterns (Woodruff, Huebner & Telfer, 1986), whilst, in some species, those overlying the oocyte may contribute to the accumulation of yolk precursors in the intercellular space (Kelly & Telfer, 1979; Telfer, Huebner & Smith, 1982) and/or the formation of different egg envelopes. Brennan, Weiner, Goralski & Mahowald (1982) demonstrated with in situ hybridization that the follicle cells of Drosophila are a major ovarian site of yolk protein synthesis. The temporal evolution of ovarian yolk polypeptide synthesis in Drosophila has been studied by Isaac & Bownes (1982), but no information was provided about the spatial pattern of this synthesis. In this paper, we focus on the distribution of yolk polypeptides in the follicle cells during their differentiation. In addition to this, we present evidence for a correlation between general morphology, the degree of polyploidy, the rate of messenger RNA synthesis and the presence (or absence) of yolk proteins in the follicle cells.
Materials and methods
Morphology and stage determination
S. bullata was reared in crowded conditions as described by Huybrechts & De Loof (1977). Ovaries were dissected in a balanced Ringer solution (Chan & Gehring. 1971) and immersed in Bouin–Hollandes fixative supplemented with 10 % of a saturated aqueous mercury chloride solution for 24h, rinsed in tap water for 24h, dehydrated through standard series of graded ethanols, ethanol-xylol and xylol. Finally, ovaries were embedded in Paraplast and sectioned at 5 μm. Vitellogenic stages were classified according to the relative ratio of oocyte length over follicle length (modified classification of Pappas & Fraenkel, 1977) and by means of cytological features. Vitellogenic stages 4A, 4B and 4C will be referred to as early, mid and late vitellogenesis.
Electrophoresis
Ringer and haemolymph were drained from dissected ovaries, which were then pooled according to vitellogenic stage in Eppendorf tubes and weighed. Haemolymph was collected from a wound near the mesothoracic leg in microcapillaries, which were blown out on an aluminium block cooled to liquid nitrogen temperature. This method allows collection of haemolymph without antioxidant additives. 9 μl of SDS sample buffer were added per mg of sample (wet weight for ovaries). Samples were simultaneously sonicated and heated to boiling temperature in a Soniprep 150 sonicator. Solid materials were sedimented in an Eppendorf centrifuge and supernatants stored at −20°C. SDS–PAGE was performed using 5–15 % gradient gels with the Laemmli (1970) buffer system. Gels were stained by means of Coomassie Brilliant Blue R250.
Antiserum preparation
Yolk polypeptides from an equivalent of 5 mg of egg homogenate (Huybrechts & De Loof, 1982), purified by means of SDS–PAGE, were injected into rabbits. Three similar booster injections were given biweekly, starting three weeks after the first injection. Animals were terminally bled from the dorsal aorta one week after the fourth injection.
Immunoblotting, ‘Aurodye’ staining and immunocytochemistry
Protein bands were transferred from SDS gels to nitrocellulose by means of isotachophoresis (Kyhse-Andersen, 1985) using a self-made semidry blotting apparatus and visualized with Aurodye staining (Janssen Life Science products, Beerse, Belgium) (Moeremans, Daneels, de Raeymaeker & De Mey, 1987) or Amido Black 10B. Yolk proteins were visualized by means of peroxidase antiperoxidase (PAP) immunostaining for blotted antigens (Geysen, Vandesande & De Loof, 1984; Geysen, 1987) and for tissue section antigens (Vandesande, 1983; Verhaert, Geysen, De Loof & Vandesande, 1984). For both techniques, the optimal dilution of anti-yolk protein antiserum proved to be 1/10000.
Solid-phase immunoadsorption and antiserum specificity
Haemolymph samples of 5-day-old females and males were linked to cyanogen-bromide-activated Sepharose (Pharmacia) as specified by the manufacturers. 2 ml of antiserum diluted 1/10000 was incubated three times for 2h with 5 μl equivalents of either male or female immobilized haemolymph. Immunoadsorbed sera were then applied to tissue sections in the PAP technique. Immunostaining could be abolished completely with female haemolymph whereas serum treated with male haemolymph reacted as the native antiserum (Fig. 1A,B,C). Because yolk proteins are female-specific, this assay indirectly proves serum specificity. Immunoblotting revealed that in ovarian preparations no protein other than yolk proteins are recognized by the antiserum (Fig. 1D).
Specificity of the antibody in PAP immunostaining using immunoadsorption (A,B,C) and immunoblotting (D). Mid-vitellogenic ovary stained with untreated antiserum (A), after immunoadsorption with immobilized male (B) and female (C) haemolymph. Immunostaining is abolished after adsorption with female haemolymph only. (D) Immunoblotting. Female and male haemolymph in lanes 1 and 2, respectively; ovarium homogenate in lane 3. Lanes 4 and 5: strips of a gel overloaded with sample to check for minor contaminants to ovarial products. Aurodye-stained strip (4) PAP-immunostained strip (5). The antibody is specific for the yolk polypeptides.
Specificity of the antibody in PAP immunostaining using immunoadsorption (A,B,C) and immunoblotting (D). Mid-vitellogenic ovary stained with untreated antiserum (A), after immunoadsorption with immobilized male (B) and female (C) haemolymph. Immunostaining is abolished after adsorption with female haemolymph only. (D) Immunoblotting. Female and male haemolymph in lanes 1 and 2, respectively; ovarium homogenate in lane 3. Lanes 4 and 5: strips of a gel overloaded with sample to check for minor contaminants to ovarial products. Aurodye-stained strip (4) PAP-immunostained strip (5). The antibody is specific for the yolk polypeptides.
Poly(A)+ RNA extraction
Previtellogenic, vitellogenic and maturing ovaries were dissected and frozen immediately in liquid nitrogen. Total RNA was extracted as described by Cardoen, Huybrechts & De Loof (1986a). Poly(A)+ RNA fractions were isolated by oligo(dT) chromatography (Aviv & Leder, 1972), 2·7 % to 3·0 % of total RNA proved to be poly(A)+ RNA. After precipitation with two volumes of ethanol and 0·3 M-sodium acetate, fractions were stored at −20°C.
Cell-free translation, fluorography and immunoprecipitation
The poly(A)+ RNA was washed twice with 70% ethanol and finally dissolved in sterile deionized water at 1 μg μl−1. 3ug of poly(A)+ RNA was translated in 40 μl of rabbit reticulocyte lysate mixture (Amersham, N90) in the presence of 60 μCi [35S]methionine (Amersham) for 1 h at 30°C. Subsequently, the samples were counted and the translation products were analysed by SDS–PAGE using 8–12% gradient gels (Laemmli, 1970). About 100000cts min-1 per lane were applied. After fixation, the gel was treated with 1 M-sodium salicylate according to Chamberlain (1979), dried and finally exposed to a Kodak X-omat film at −70°C. For immunoprecipitation, 5 μl of reticulocyte lysate was incubated with 5 μl of undiluted antiserum for 30 min at 37°C and overnight at 4°C. Secondary antibody was added and a similar incubation schedule followed. Immunoprecipitates were then sedimented using an Eppendorf centrifuge, washed and resuspended in Trisbuffered saline and precipitated (five cycles).
Total RNA synthesis
For monitoring RNA synthesis and transport, females were injected with 6 μCi [5,63H]uridine (Amersham, 45 Ci mmol−1). Ovaries were dissected in Carnoy fixative, dehydrated, embedded in Paraplast and sectioned at 4 μm. Control flies were injected with 6 μl of a 0·1 M solution of cold uridine (Serva); the specificity of the [3H]uridine incorporation was checked by including a RNAse digestion before application of the stripping film (Ray & Ramamurty, 1979).
In situ hybridization with [3Hjpoly(U)
In order to investigate the distribution of poly(A)+ RNA ovarian tissue sections were hybridized using a [3H]poly(U) probe (Amersham, 41–147 residues; 20–72 Ci mmol−1) as a homopolymer probe. Hybridization and autoradiography were performed as described by Cardoen et al. (1986b).
Results and discussion
Previtellogenic stages
In all previtellogenic stages, the follicle cells are immunonegative for yolk polypeptides. At the onset of vitellogenesis, the fat body, haemolymph and peripheral ooplasm are immunopositive, due to the synthesis, transport and endocytotic uptake of exogenous yolk, but immunostaining is absent in follicle cells (Fig. 2A–C). At the anterior pole, the proliferation of the border cell complex indicates the first morphological differentiation of follicle cells (Fig. 2B).
PAP immunostaining for yolk protein; Normarski differential interference contrast. (A) First uptake of exogenous yolk in the cortical ooplasm (white asterisks); all follicle cells are negative, haemolymph positive (black asterisk). (B) Detail of border cell proliferation. (C) Follicle cells covering the oocyte. (D) Early vitellogenesis (stage 4A); ooplasm and all follicle cells positive for yolk proteins. (E) Detail of the transition zone from squamous to columnar follicle cells. More squamous cells are poorly labelled (top), columnar cells display an immunopositive granulation apically and basally from the nuclei (bottom). Bars in all photographs, 20 μm.
PAP immunostaining for yolk protein; Normarski differential interference contrast. (A) First uptake of exogenous yolk in the cortical ooplasm (white asterisks); all follicle cells are negative, haemolymph positive (black asterisk). (B) Detail of border cell proliferation. (C) Follicle cells covering the oocyte. (D) Early vitellogenesis (stage 4A); ooplasm and all follicle cells positive for yolk proteins. (E) Detail of the transition zone from squamous to columnar follicle cells. More squamous cells are poorly labelled (top), columnar cells display an immunopositive granulation apically and basally from the nuclei (bottom). Bars in all photographs, 20 μm.
The genes coding for yolk polypeptides in the fat body and the follicle cells seem to respond to the same hormonal environment in a different way. The time lag may either be due to different susceptibilities of the two tissues or to the action of different triggers as has been suggested for Drosophila by Isaac & Bownes (1982). Regulation of follicle cell contribution to vitellogenesis seems to be a complex mechanism. An explanation for tissue specificity of hormonal action has been described previously (De Loof, 1986û).
Vitellogenic stages
In early vitellogenic egg follicles, the follicular epithelium differentiates into distinct zones (Fig. 2D); (1) a border cell complex at the anterior end of the follicle, (2) an adjacent squamous epithelium, (3) a transition zone with follicle cells of intermediate cytological appearance (from squamous to columnar) and (4) columnar cells overlying the oocyte surface, which are highest at the nurse cell-oocyte border, and somewhat more cuboidal near the posterior end (Fig. 2D,E). At this stage, all follicle cells are immunopositive for yolk polypeptides, but the stained material is not equally distributed. The cytoplasm of both the border cells and the squamous epithelium reacts weakly, which contrasts with the complete absence of staining in the nurse cells. Posteriorwards from the transition zone, the immunopositivity increases to a maximum at the nurse cell-oocyte border and in the epithelium covering the oocyte. Basally and apically from the follicle cell nuclei, very dense deposits of immunostaining occur, indicating that yolk polypeptides are more concentrated at certain sites in the follicle cell cytoplasm (Fig. 2E). These deposits most probably correspond to large stacks of rough endoplasmic reticulum (RER) located basally and to RER or the Golgi system present in the apical cytoplasm of these follicle cells (Geysen, Cardoen, Huybrechts & De Loof, 1986).
The transition zone is remarkable because it implies that different types of follicle cells do not necessarily originate from predetermined groups of follicle cells. Regulative mechanisms may determine the differentiational pathway of a follicle cell up to mid-vitellogenesis. The pairs of polar cells at each pole of the follicle (Geysen & De Loof, 1983) may play a role as organizing centres.
At mid-vitellogenesis (Fig. 3B–D; Fig. 3A shows an intermediary stage), the nurse cells are covered by squamous follicle cells, whereas the columnar epithelium overlies the oocyte chamber and occupies the oocyte-nurse cell border. The border cells reach a position between oocyte and nurse chamber. The transition zone is no longer present, but a sharp delineation between the two types occurs at the nurse cell–oocyte border. This differentiation is also reflected in the immunostaining pattern: only the columnar follicular epithelium is immunopositive whereas the squamous follicle cells and the border cells have lost immunoreactive material and contrast significantly with the immunopositive haemolymph and the ooplasm respectively (Fig. 3B,C,D). Apparently, these follicle cells diminish or stop synthesizing yolk polypeptides rapidly in the course of differentiation. We have indications that the squamous follicle cells contribute to cytoplasmic transport from the nurse cells to the oocyte (for review see Gutzeit. 1986) as they contain considerable amounts of microfilaments (Geysen & De Loof, 1986).
PAP immunostaining for yolk protein; Nomarski differential interference contrast. (A) Intermediate stage between 4A and 4B. Start of border cell migration (arrowhead), the transition zone shifts towards the oocyte-nurse cell border; border cells and squamous follicle cells lose yolk protein immunoreactivity. (B) Mid-vitellogenesis (stage 4B). border cells (arrowhead) reach the oocyte and squamous follicle cells cover the nurse cell chamber. The transition zone has turned into a sharp delineation between positive and negative cells (double arrowhead). (C) Detail of two immunonegative squamous follicle cells (arrowheads), contrasting with immunopositive haemolymph (asterisk).(D) Detail of oocyte-nurse cell border. Immunopositive follicle cells display granulation apically and basally from the nuclei; squamous cells (arrowhead).
PAP immunostaining for yolk protein; Nomarski differential interference contrast. (A) Intermediate stage between 4A and 4B. Start of border cell migration (arrowhead), the transition zone shifts towards the oocyte-nurse cell border; border cells and squamous follicle cells lose yolk protein immunoreactivity. (B) Mid-vitellogenesis (stage 4B). border cells (arrowhead) reach the oocyte and squamous follicle cells cover the nurse cell chamber. The transition zone has turned into a sharp delineation between positive and negative cells (double arrowhead). (C) Detail of two immunonegative squamous follicle cells (arrowheads), contrasting with immunopositive haemolymph (asterisk).(D) Detail of oocyte-nurse cell border. Immunopositive follicle cells display granulation apically and basally from the nuclei; squamous cells (arrowhead).
From mid to late vitellogenesis, the immunopositive follicle cells spread and change from columnar to cuboidal and finally squamoid shape. Immunopositive granules in the cytoplasm predominantly appear laterally from the oocyte. The immunonegative vitelline membrane separates oocyte and follicle cells (Fig. 4B,C). At the oocyte–nurse cell border, the follicle cells move centripetally and incompletely separate both compartments. The immunopositive follicle cells and the squamous epithelium perform this centripetal movement in different ways: cell bodies and nuclei of the latter can be observed in centripetal position, whereas from the immunopositive follicle cells only a thin layer of cytoplasm is present centripetally, but no cell bodies or nuclei (Fig. 4A). After the squamous cells have moved in deeply centripetally, yolk-polypeptide-positive cell bodies and nuclei can be observed in a more centripetal position.
PAP immunostaining for yolk proteins; Nomarski differential interference except B and E: transillumination. (A) Stage 4C. Centripetal movement. Squamous cells (arrowheads) migrate first whereas positive follicle cells seem to send cytoplasmic projections along (double arrowhead); haemolymph (asterisk) stains weakly. (B.C) Positive follicle cells spread over the growing oocyte surface. Positive granules now are mainly located laterally from the nuclei. The immunonegative vitelline membrane (arrowhead) separates follicle cells and oolemma. (D) Mature follicle (stage M), nurse cell degeneration is completed. All follicle cells are negative, border cells proliferate (arrowhead); egg shells are deposited (double arrowhead). (E,F) The stretched follicle cells are fully immunonegative (E) but still display clear nuclear morphology (F). Arrowheads, nucleoli.
PAP immunostaining for yolk proteins; Nomarski differential interference except B and E: transillumination. (A) Stage 4C. Centripetal movement. Squamous cells (arrowheads) migrate first whereas positive follicle cells seem to send cytoplasmic projections along (double arrowhead); haemolymph (asterisk) stains weakly. (B.C) Positive follicle cells spread over the growing oocyte surface. Positive granules now are mainly located laterally from the nuclei. The immunonegative vitelline membrane (arrowhead) separates follicle cells and oolemma. (D) Mature follicle (stage M), nurse cell degeneration is completed. All follicle cells are negative, border cells proliferate (arrowhead); egg shells are deposited (double arrowhead). (E,F) The stretched follicle cells are fully immunonegative (E) but still display clear nuclear morphology (F). Arrowheads, nucleoli.
At late vitellogenesis, the intensely stained granules have disappeared from the follicle cells covering the oocyte surface, but a weak cytoplasmic immunostaining is still present.
Postvitellogenic stages
The follicle cells become completely immunonegative for yolk polypeptides only at the end of nurse cell degeneration. However, differential interference microscopy reveals that the follicle cells covering the oocyte surface have not yet disintegrated (Fig. 4E.F). They probably make a functional switch from yolk polypeptide to chorion protein synthesis. Morphologically, the border cells do not change significantly throughout vitellogenesis, although transcriptionally they are as active as the follicle cells covering the oocyte (Cardoen et al. 1986b). During and after nurse cells degeneration they proliferate and form the micropyle. Also, the proliferation of the squamous epithelium continues during and after nurse cell breakdown (Fig. 4D).
Nuclear morphology and transcriptional activity of the follicle cells
The nuclei of the follicle cells adjacent to the oocyte are round while those of the follicle cells surrounding the nurse cells are flattened (Fig. 5A). This morphological differentiation entails differences in DNA content (Cardoen et al. 1986b). Time-dependent labelling experiments with [3H]uridine (Fig. 5B) suggests that the columnar follicle cells are more actively involved in RNA synthesis than the squamous ones. In situ hybridization with [3H]poly(U) suggests an intensive transcriptional activity in the follicle cells covering the oocyte and in the border cells (Fig. 5C) (see also Cardoen et al. 19866). Although these differences may also be due to different RNA-túrnover rates, preservation of RNA in the section or accessibility to the probe, these effects reflect aspects of the divergent differen-tiational pathways of the two follicle cell types.
Nuclear morphology, polyploidy and transcriptional activity. (A) Feulgen staining of stage-4B vitellogenic follicles; sections were hydrolysed in 4N-HC1 for 50min prior to staining (Cardoen et al. 19866). Nuclear morphology (and polyploidy) of squamous follicle cells (arrowheads) differs from that of the columnar follicle cells and the border cells (double arrowhead). (B) In vivo labelling with [3H]uridine for 1 h. The columnar follicle cells (double arrowhead) incorporate more label than the squamous in nucleus and cytoplasm. There is no evidence for mRNA transport from follicle cells into the oocyte (C). In situ hybridization with labelled poly(U)+ RNA. mRNA synthesis is intense in the follicle cells overlaying the oocyte (double arrowhead) and low in squamous follicle cells (single arrowheads). The karyosome also displays weak activity (triple arrow’head). (D) Fluorography of in vitro translated mRNA fractions from: lane 1. previtellogenic ovaries; lane 2, vitellogenic ovaries; lane 3, fat body from vitellogenic females; lane 4, mature ovaries; lane 5, female haemolymph, labelled in vivo by injection of tritiated amino acids. Due to processing of yolk proteins by the fat body, the yolk proteins from haemolymph (lane 5) have a smaller molecular weight than unprocessed yolk polypeptides. (E) Immunoprecipitation with anti-yolk protein antiserum of translated mRNA from: lane 1, mature ovaries; lane 2, previtellogenic ovaries; lane 3, fat body of vitellogenic females; lane 4. vitellogenic ovaries. Labelled yolk proteins appear in. and were only immunoprecipitated from, vitellogenic ovaries and fat body (weak reaction) of vitellogenic females.
Nuclear morphology, polyploidy and transcriptional activity. (A) Feulgen staining of stage-4B vitellogenic follicles; sections were hydrolysed in 4N-HC1 for 50min prior to staining (Cardoen et al. 19866). Nuclear morphology (and polyploidy) of squamous follicle cells (arrowheads) differs from that of the columnar follicle cells and the border cells (double arrowhead). (B) In vivo labelling with [3H]uridine for 1 h. The columnar follicle cells (double arrowhead) incorporate more label than the squamous in nucleus and cytoplasm. There is no evidence for mRNA transport from follicle cells into the oocyte (C). In situ hybridization with labelled poly(U)+ RNA. mRNA synthesis is intense in the follicle cells overlaying the oocyte (double arrowhead) and low in squamous follicle cells (single arrowheads). The karyosome also displays weak activity (triple arrow’head). (D) Fluorography of in vitro translated mRNA fractions from: lane 1. previtellogenic ovaries; lane 2, vitellogenic ovaries; lane 3, fat body from vitellogenic females; lane 4, mature ovaries; lane 5, female haemolymph, labelled in vivo by injection of tritiated amino acids. Due to processing of yolk proteins by the fat body, the yolk proteins from haemolymph (lane 5) have a smaller molecular weight than unprocessed yolk polypeptides. (E) Immunoprecipitation with anti-yolk protein antiserum of translated mRNA from: lane 1, mature ovaries; lane 2, previtellogenic ovaries; lane 3, fat body of vitellogenic females; lane 4. vitellogenic ovaries. Labelled yolk proteins appear in. and were only immunoprecipitated from, vitellogenic ovaries and fat body (weak reaction) of vitellogenic females.
Fat body and ovaries of vitellogenic females, but not pre- and postvitellogenic ovaries, contain yolk polypeptide mRNA (Fig. 5D). In fat body extracts of vitellogenic females, yolk polypeptides are the predominant transcripts, whereas in vitellogenic ovaries other messages are also transcribed. Postvitellogenic ovaries display a different transcriptional pattern. We observe small differences in molecular weight between the translated yolk polypeptides (lanes 2, 3) and the yolk polypeptides occurring in the haemolymph (lane 5). This difference is due to cleavage of a signal peptide (Cardoen et al. 1986a).
Synthesis of yolk polypeptides by the follicle cells
The follicle cells of S. bullata contain and perhaps synthesize the ovarian yolk proteins (Huybrechts. Cardoen & De Loof, 1983). The correlation between immunopositivity of the follicle cells and the presence of yolk polypeptide mRNA strongly suggests that the follicle cells are involved in yolk polypeptide synthesis during vitellogenesis. It is very unlikely that the yolk polypeptides would be synthesized in other ovarian cell types. The karyosome of the germinal vesicle exhibits only very weak transcriptional activity (Cardoen et al. 19866), the nurse cells do not display immunoreactivity and transport of RNA from the nurse cells does not occur until late vitellogenesis. We never obtained evidence indicating RNA transfer from the follicle cells to the oocyte.
It is more likely that the RNA synthesis in the follicle cells sustains translational activity. Yolk polypeptide messengers have a high turnover rate and a concomitant short life time, which would make transport inefficient and unlikely (Cardoen, Huybrechts, Theunis & De Loof, 1984; Cardoen et al. 1986a). The presence of a signal peptide is compatible with the fact that the yolk proteins are synthesized in cells possessing a well-developed RER and Golgi complex. Compared to the fat body of vitellogenic females, however, yolk polypeptide mRNA constitute only a minor part of the ovarian transcripts. The fact that the vitelline membrane is formed while the cylindrical follicle cells are immunopositive points to the complex pattern of functions these follicle cells perform.
Polar follicle cell differentiation and extracellular current patterns
Superimposed on polarized morphological differentiation, we observed four other features in Sarcophaga follicles: (1) the distinct nuclear morphology, (2) the different degrees of polyploidy (squamous cells 4C; columnar cells 8 to 16C; Cardoen et al. 1986b), (3) the different levels of transcriptional activity and (4) restriction of yolk polypeptide synthesis to the cuboidal follicle cells overlying the oocyte.
The extracellular current patterns around polytrophic follicles can be interpreted as a fifth element in polar differentiation. They do not complement the intracellular potential difference between oocyte and nurse cells but follow follicle cell differentiation in Sarcophaga (Verachtert & De Loof, 1986). This critically suggests that the follicle cells behave as an electrically independent system from the nurse cell–oocyte syncytium in the course of their differentiation. the fact that yolk proteins are synthesized in follicle cells of columnar to cuboidal shape with round nuclei and that squamification of cells is followed by a loss of the ability to synthesize yolk proteins can be interpreted following the concept of intracellular electrophoresis by cells (De Loof, 1983, 1985) and the concept of epigenetic regulation of gene expression by ions (De Loof & Geysen, 1983; De Loof, 19866). The well-known morphology and new insights into the physiology of the polytrophic follicle make it a favourable model to elaborate epigenetic aspects of gene expression.
ACKNOWLEDGEMENTS
The authors are grateful to Prof. Spencer Berry, Dr Roger Huybrechts and Dr Barend Verachtert for critically reading the manuscript. We also wish to thank Prof. Frans Ollevier for allowing access to the Reichert Polyvar microscope and Julie Puttemans for photographical assistance. J.C. wishes to thank the IWONL (Belgium) for financial support.