Origin and spatial distribution of maternal messenger rna during oogenesis of an insect, Oncopeltus fasciatus

The developmental patterns typical of mosaic eggs may be caused by morphogenetic substances localized in the egg cytoplasm (Wilson, 1925; Davidson, 1976). Although the biochemical nature of these substances is unknown, the most likely candidates appear to be informational macromolecules such as maternal mRNA and proteins (Whittaker, 1973, 1977; Kandler-Singer & Kalthoff, 1976; Jeffery & Capco, 1978; Rodgers & Gross, 1978). If maternal mRNA and proteins are involved in the localization phenomena a differential distribution of these molecules in the egg cytoplasm might be anticipated. In order to investigate the spatial distribution of maternal mRNA in eggs we have developed a method for the detection of poly(A)-containing RNA [poly(A) + RNA], a class of mRNA (Brawerman, 1974), in histological sections by in situ hybridization with [³ H]-poly(U) (Capco & Jeffery, 1978). Applying this method to the eggs of the milkweed bug, Oncopeltus fasciatus, we observed a uniform distribution of poly(A) + RNA molecules in the cytoplasm prior to syncytial blastoderm formation. In the present study we have extended our original application of the [³ H]-poly(U) in situ hybridization technique to Oncopeltus ovaries. Our objective was to investigate the origin and spatial distribution of maternal mRNA during oogenesis.

The meroistic ovaries of insects offer unique material to study the origin and depositional pattern of maternal mRNA. Two types of accessory cells, the trophocytes and the follicle cells, as well as the oocyte genome itself are potential sites of transcription. In this type of ovary direct cytoplasmic connexions, which have been shown to function in inter-cellular transport of RNA (King, 1960; Mahowald, 1973), exist between the developing oocytes and the trophocytes. In this report we present evidence which suggests that maternal mRNA molecules originate from at least 2 sources during oogenesis and that these molecules are subject to quantitative localization in the polar cytoplasms of late vitellogenic oocytes.

Cultures of Oncopeltus fasciatus were raised in plastic containers on a 14 h: 10 h, light:dark cycle as described by Beck, Edwards & Medler (1958). Ovaries were dissected from animals in peak oviposition stage, about 2 weeks after their final moult. The ovarioles were separated and fixed in absolute ethanol:acetic acid (3: 1) for 15 min. The fixed ovarioles were dehydrated through a graded ethanol series, embedded in paraplast, and sectioned at 10 μm. In situ hybridization with [³H]-poly(U) was carried out according to the procedure of Capco & Jeffery (1978) except that the concentration of [³H]-poly(U) applied to the sections was 1·8 μCi/ml (4·65 mCi/M; New England Nuclear Corp., Boston, Ma.). Briefly, slides were pretreated with 100 μg/ml DNase dissolved in 100 mM Tris-HCl (pH 7·6)–3 mM MgCl₂ for 1 h at 37 °C, annealing was performed in 10 mM Tris-HCl (pH 7·6)-200 MM NaCl-5 mM MgCl₂ for 3 h at 50 °C, and after annealing the slides were successively rinsed with the hybridization buffer and 50 MM Tris-HCl (pH 7·6)–100 mM KCl-1 mM MgCl₃ (TKM). Uncomplexed [³H]-poly(U) was hydrolysed by treatment of the slides with 50 μg/ml pancreatic ribonuclease (RNase) A dissolved in TKM for 1 h at 37 °C and removed by extraction with cold 5 % trichloroacetic acid for 10 min. Extracted slides were rinsed in distilled water, air-dried, and autoradiographed using Kodak NTB-2 liquid emulsion (Eastman-Kodak Special Products Division, Rochester, N.Y.). Autoradiographs were exposed for 14 days, developed, and stained through the emulsion with Harris haematoxylin-eosin.

Fig. 1 shows a longitudinal section through 2 Oncopeltus ovarioles subjected to in situ hybridization with [³H]-poly(U). The dark areas on this low-magnification micrograph represent regions of intense labelling. The trophocytes at the anterior tip of the germarium and the follicle cells surrounding the oocytes are heavily labelled while the oocytes themselves exhibit less-intense labelling. The [³H]-poly(U) binding sites present in these ovariole sections, like those of Oncopeltus eggs and embryos (Capco & Jeffery, 1978), are selectively removed by pretreatment with RNase A dissolved in low-ionic-strength buffers suggesting they represent poly(A) sequences (Beers, 1960; Darnell, Wall & Tuchinski, 1971).

The distribution of grains seen in the germarium following [³H]-poly(U) in situ hybridization is shown in Figs. 2 –5 and quantified in Table 1. Nuclear labelling appeared to be constant in trophocytes of all stages. However, the cytoplasmic grains were more concentrated in mature zone III trophocytes (see Bonhag, 1955, for trophocyte staging) than in zone II trophocytes and the latter, in turn, exhibited more intense cytoplasmic labelling than zone I cells (Figs. 2 –4). The overall increase in cytoplasmic labelling observed in zone III relative to zone I trophocytes was about 4-f0ld. These results suggest that poly(A) + RNA accumulates in the cytoplasm during trophocyte maturation.

The trophic cores and nutritive tubes were also labelled following in situ hybridization, but their grain concentration was less than 30% of that found in the cytoplasm of the zone III trophocytes (Fig. 5; Table 1). Since nutritive tubes function in the transport of trophocyte-derived products to the growing oocytes (Bonhag, 1955; MacGregor & Stebbings, 1970; Zinsmeister & Davenport, 1971; Davenport, 1976) the presence of poly(U)-binding sites in these structures implies that poly(A)+RNA may be among the transported substances. The low level of poly(U)-binding activity in the trophic cores and nutritive tubes relative to the zone III trophocyte cytoplasm could mean that poly(A) + RNA transport through these structures is very rapid or that the entire population of trophocyte poly(A) + RNA is not donated to the oocyte. The limiting factor in poly(A) + RNA transport may be related to the density of cytoskeletal binding sites in the nutritive tubes.

As shown in Fig. 6, stage 2 oocytes (see Schreiner, 1977, for oocyte staging), the youngest analysed, exhibited low levels of cytoplasmic labelling compared to that found in the trophocytes, trophic cores, or nutritive tubes. In more mature oocytes the grain concentration per unit endoplasmic area gradually increased until a maximum was reached at stage 5. Afterwards a steady decline in grain density began (Fig. 6). The low grain concentrations seen in stage 7 oocytes approximated those previously observed in freshly oviposited Oncopeltus eggs (Capco & Jeffery, 1978).

Since Oncopeltus oocytes increase markedly in volume during oogenesis due to growth and yolk addition, the grain concentration per unit endoplasmic area is not representative of the actual change in poly(U)-binding site number per oocyte. When total oocyte labelling was approximated, by multiplying the endoplasmic grain number of each total oocyte section by the estimated oocyte volume, it was found that the [³H]-poly(U)-binding site titre per oocyte continually increased during oogenesis, with the largest increases seen between stages 6’ and 7 (Fig. 7). The increase in [³H]-poly(U)-binding sites which occurs after the nutritive tubes have ceased transport of materials into the oocyte (Bonhag, 1955) suggests that post-vitellogenic oocytes accumulate poly(A) + RNA molecules which originate from a source other than the trophocytes.

In contrast to the oocyte cytoplasm, none of the nuclei examined during early oogenesis (stage 2–4 oocytes) showed significant labelling (Fig. 12). These results suggest that, at least in pre-vitellogenic and early vitellogenic oocytes, the oocyte genome may not be active in poly(A) + RNA formation.

The oocyte cytoplasmic regions were also scrutinized for differences in labelling following [³ H]-poly(U) in situ hybridization. In young (stage 2–5) oocytes grains were uniformly distributed throughout the cytoplasm. However, when the endoplasmic grain concentration began to be diluted by oocyte growth (Figs. 6–8), labelling in the anterior and posterior polar periplasms did not decrease and in some cases was intensified (Figs. 9–10; Table 2). The lateral periplasmic regions, unlike those at the oocyte poles, did not usually exhibit grain concentrations significantly different from the endoplasm (Fig. 11; Table 2). Although considerable variation in labelling occurred between cytoplasmic regions of oocytes from different ovarioles, each stage 6″ oocyte examined showed about a 2-fold and sometimes a much greater enhancement of labelling in the polar periplasms relative to the endoplasm (Table 2). The variability of labelling seen in the polar periplasms of different ovarioles may be due to asynchronous growth of oocytes classified as the same developmental stage.

These results are consistent with the possibility that poly(A)+RNA molecules are enriched in the polar regions of the oocyte during late vitellogenesis. The polar accumulations could not be detected in stage 7 oocytes.

The follicle cell cytoplasm was also intensely labelled following in situ hybridization with [³H]-poly(U) (Figs. 13–18). Labelling was particularly pronounced in the cytoplasmic region adjacent to the oocyte surface at stages 6′ and 6″ (Figs. 13, 15). At this time pseudopod-like projections, which have been previously described in the telotrophic meroistic ovaries of Gerris (Eschenberg & Dunlap, 1966) and Rhodnius (Huebner & Anderson, 1972), were observed (Fig. 16). These projections sometimes appear as heavily labelled periplasmic spheres (Fig. 17), presumably due to sectioning through their knob-like termini. After chorion formation the projections disappeared but the follicle cell cytoplasm remained labelled (Fig. 18).

In the meroistic ovaries of insects transcription of maternal mRNA, which is eventually deposited in the oocyte, could occur in the trophocytes, the follicle cells, or the oocyte nucleus itself (Mahowald, 1973). The present results, which have been obtained by the use of in situ hybridization with [³H]-poly(U) for the detection of poly(A) + RNA (Capco & Jeffery, 1978; Jeffery & Capco, 1978), suggest that at least 2 of these potential transcript sources donate maternal poly(A) + RNA to the Oncopeltus oocyte. Furthermore, a localized accumulation of maternal poly(A) + RNA has been detected in the polar periplasmic regions of late vitellogenic oocytes.

During early oogenesis the trophocytes are probably the major source of oocyte poly(A)+RNA. Since the in situ hybridization procedure we have employed cannot provide direct information on the transport of poly(A) + RNA, evidence for the trophocyte origin of these molecules is based on the presence of poly(U)-binding sites in the nutritive tubes, microtubule-containing corridors known to participate in RNA transfer (Bier, 1963; Vanderberg, 1963; MacGregor & Stebbings, 1970; Zinsmeister & Davenport, 1971; Davenport, 1974, 1976). However, the possible contribution of poly(A) + RNA by follicle cells which border these passages cannot be excluded by these data. Additional evidence for an extra-oocyte origin of early oogenetic maternal poly(A) + RNA is provided by the lack of [³H]-poly(U)-binding sites in the nuclei of stage 2∼4 oocytes. The postulated trophocyte contribution of oocyte poly(A) + RNA is also consistent with our results showing that poly(A) + RNA accumulates in the cytoplasm of maturing trophocytes and the conclusions of radio-nuclide incorporation studies recently carried out with silk moth ovaries (Paglia, Berry & Kastern, 1976).

Following ligature of the nutritive tubes of Oncopeltus ovaries, Davenport (1976) detected very low levels of [³H]uridine incorporation into high-molecular-weight, heterodisperse oocyte RNA and concluded that the oocyte nucleus may participate in maternal mRNA transcription during early oogenesis. Our results are not entirely incompatible with the previous work since Davenport did not establish that RNA synthesis actually occurred in the oocyte nucleus. Thus, the transcriptional activity observed by Davenport (1976) could possibly be of mitochondrial origin. The ligation itself could also induce nuclear transcription in the oocyte. Alternatively, the absence of appreciable [³ H]-poly(U)-binding sites in the oocyte nucleus uncovered by our studies could also be explained by the masking of nuclear poly(A) sequences by proteins (Kwan & Brawerman, 1972) or other polynucleotide tracts (Jeffery & Brawerman, 1975).

It is clear from the present study that at least one other source of maternal mRNA besides the trophocytes must exist since [³H]-poly(U)-binding sites markedly increase in quantity after the nutritive tubes are severed by chorion formation (Bonhag, 1955). The transcription of poly(A) + RNA in post-vitellogenic oocytes has also been described in Dysdercus (Winter, Wiemann-Weiss & Duspiva, 1977). The identity of the other site(s) of oocyte maternal mRNA transcription in post-vitellogenic insect oocytes is currently unresolved. However, it is possible that the oocyte nucleus may become transcriptionally active at this time. Unfortunately, we could not obtain favourable sections of Oncopeltus stage 7 oocyte nuclei to test this hypothesis.

The high concentration of [³ H]-ρoly(U)-binding sites observed in the follicle cell cytoplasm suggests that it is rich in poly(A) + RNA. However, it is questionable whether these molecules are transported into the oocyte as originally proposed by Bier (1963) and by Vanderberg (1963). Our current observation of intense [³ H]-poly(U)-binding activity within the follicle cell projections into the stage 6″ oocyte cytoplasm is not sufficient to suggest follicle cell RNA transfer since abnormally high concentrations of binding sites were not observed on the oocyte sides of these structures. It is more likely that the follicle cell poly(A) + RNA molecules which accumulate proximal to the surface of stage 6″ oocytes are involved in the precocious synthesis of chorion proteins. Thus, we must conclude, as did Telfer (1964) previously, that no evidence currently exists for the derivation of oocyte maternal mRNA from the follicle cells.

A major objective of the present investigation was to test for the existence of maternal mRNA localizations in Oncopeltus oocytes. Our previous study, which also employed in situ hybridization with [³H]-poly(U) for poly(A) + RNA detection, suggested that such localizations were not present in freshly oviposited eggs (Capco & Jeffery, 1978). Our current results also suggest that oocyte poly(A) + RNA molecules are uniformly distributed throughout the cytoplasm during most of oogenesis. However, one notable exception to this conclusion is the situation discovered in late vitellogenic oocytes in which accumulations of [³ H]-poly(U)-binding sites appeared in the anterior and posterior polar periplasms. If the differential distribution of these binding sites actually reflects variations in the titre of poly(A) + RNA molecules, rather than changes in relative length or accessibility of the poly(A) sequence itself, these findings imply that the polar periplasmic regions of late vitellogenic Oncopeltus oocytes are preferentially enriched in maternal mRNA.

Since the polar localizations of poly(A)+RNA present in vitellogenic oocytes could not be detected after chorion formation they may be transient phenomena. However, masking or preferential turnover of their poly(A) sequences could also explain their absence in stage 7 oocytes and freshly oviposited eggs. In any case the activity of the polar mRNA molecules could be instrumental in the construction of regional cytoplasmic potentials which are of consequence during early embryogenesis.

We wish to express our appreciation to Dr Hugh S. Forrest for provision of Oncopeltus fasciatus cultures. Support for this work has been furnished in part by grants to W.R.J. from the NSF (PCM-77-24767) and PHS (GM26119).