Nuclear and cytoplasmic localization of newly synthesized proteins in the early mouse embryo

Here we report our results from an examination of the intracellular distribution of the newly synthesized proteins at the 1-, 2- and 4-cell stage of the early mouse embryo. Nuclear or cytoplasmic localization of proteins at the relatively transcriptionally inactive pronuclear stage, at the 2-cell stage as the embryonic nuclei are becoming transcriptionally active (Flach et al. 1982) and at the transcriptionally active 4-cell stage was determined. In addition, male and female pronuclei were analysed separately in order to determine whether the functional differences between parental genomes (Surani et al. 1984; McGrath & Solter, 1984) are associated with qualitative differences in pronuclear proteins.

Studies on amphibian oocytes have demonstrated that in addition to histones (Adamson & Woodland, 1974; Merriam & Hill, 1976), other proteins such as actin (DeRobertis et al. 1983), nucleoplasmin (Krohne & Franke, 1980; Mills et al. 1980; Laskey & Earnshaw, 1980) and so-called N proteins (Bonner, 1975) are highly concentrated within the germinal vesicle (GV) as compared to the cytoplasm. Furthermore, a large number of GV-associated proteins are transferred to the embryonic nuclei at some time during embryogenesis (Dreyer et al. 1983). In the mouse oocyte, a highly phosphorylated 26x10³A/_r (26 K) protein has been reported to be GV-associated (Wassarman et al. 1979).

The changing pattern of protein synthesis through the early preimplantation stages of mouse development has been established (Levinson et al. 1978; Howlett & Bolton, 1985 + refs therein). Only a few proteins such as histones (Wassarman & Mrozak, 1981), ribosomal proteins (LaMarca & Wassarman, 1979), tubulin (Schultz et al. 1979) and actin (Van Blerkom & Manes, 1974) have so far been identified. A previous attempt to localize newly synthesized proteins in the early mouse embryo reported no clear differences between ‘cytoplasmic’ and ‘nuclear’ fractions (Howe & Solter, 1979). Here we have examined newly synthesized nuclear and cytoplasmic proteins in the 1-, 2- and 4-cell embryo after extraction of nuclei by micromanipulation. This approach revealed very distinct differences between the new translation products that remained cytoplasmic and those that became associated with the nucleus during the early embryonic cell cycles.

Female (C57BL/6J × CBA/Ca)F₁ (AFRC colony from Bantin & Kingman stock) were superovulated by injection of 7·5i.u. pregnant mare’s serum (PMS, Intervet). 48h later these mice were injected with 7·5i.u. of human chorionic gonadotrophin (hCG, Intervet) and mated overnight with CFLP males (AFRC colony from Bantin & Kingman stock).

Fertilized eggs were recovered from plug-positive females at about 17 h post-hCG. Cumulus cells were removed by incubation in 300i.u.ml⁻¹ hyaluronidase (Sigma) in phosphate-buffered medium (PBI; Whittingham & Wales, 1969) plus 4 mg ml⁻¹ bovine serum albumin (BSA; Sigma) for 1−2 min. Eggs were then washed through six drops of T6 medium (Howlett et al. 1987) plus 4 mg ml¹ BSA and cultured under paraffin oil (BDH) in this medium at 37·8°C in humidified 5 % CO₂ in air.

Embryos were labelled for 3 h as 1-, 2- or 4-cell embryos in T6 + BSA containing [³⁵S]methionine (1100 Ci mmol⁻¹, Amersham) at a concentration of 200 μCiml⁻¹, washed through three drops of T6 + BSA and cultured for a further 2h before being prepared for micromanipulation.

Labelled embryos were placed into PBI + BSA containing 1 μg ml⁻¹ cytochalasin D (Sigma) and 0·05 μg ml⁻¹ nocod-azole (Sigma) each diluted from stocks dissolved in dimethyl sulphoxide. After 15−30min culture at 37°C, the eggs were placed in hanging drops of the same medium for micromanipulation which was carried out using a Leitz micromanipulator as described previously (Surani & Barton, 1983; Barton et al. 1984). Nuclei were extracted as karyoplasts with each nucleus surrounded by a thin film (less than 1 pm thick) of cytoplasm (see Fig. 1), thus making the cytoplasmic contamination of each karyoplast negligible (see Results).

Following manipulation to prepare karyoplasts and cytoplasts, groups of ten nuclei, individual cytoplasts or individual complete embryos were transferred to 10 pl of sample buffer (Laemmli, 1970) and analysed on 10% polyacrylamide gels exactly as described previously (Howlett & Bolton, 1985).

Fig. 1 shows groups of male and female pronuclei after removal from 1-cell eggs (Fig. 1A,B) and nuclei from 2-cell (Fig. 1C) and 4-cell (Fig. ID) embryos. There is very little obvious cytoplasmic contamination. The number and size of ‘nucleoli’ within (pro)nuclei are very variable. The diameter of male pronuclei, female pronuclei, 2- and 4-cell nuclei average about 19μm, 16μm, 14μm and 13μm, respectively. The volume of nuclear material per embryo (assuming a constant diameter of 75 μm for the total embryo) is estimated at about 1·5—3 % of the total.

Fig. 2 shows an analysis of proteins separated by onedimensional polyacrylamide gel electrophoresis. Newly synthesized proteins in intact 1-, 2- and 4-cell embryos are compared with those in nuclei and cytoplasm. Preliminary studies indicated that ten (pro)nuclei incorporated a similar level of [³⁵S]meth-ionine to that for single embryos. Hence, each sample with intact eggs and embryos (lanes A, F, I) or enucleated eggs and embryos (cytoplasts; lanes B, C, G, J) contains material from a single embryo, whereas ten (pro)nuclei were combined (lanes D, E, H, K). The experiment was repeated on four separate occasions and the patterns of proteins detected each time were consistent. It has been shown previously that cytoskeletal disruption with CCD has no gross effect on the intracellular distribution of [³⁵S]methio-nine-labelled proteins (Fey et al. 1986) and so we have assumed that the presence of CCD and nocodazole during the nuclear isolation procedure has not significantly altered the distribution of proteins.

It is apparent that there are distinct, as well as overlapping, subsets of proteins that were cytoplasmic or nuclear for each stage that was analysed. Since we did not wish to prejudge whether a labelled band at any given stage represented the same protein as at another stage, each of the three time points has been analysed and annotated separately even though some components may be the same at different stages. For each stage, nuclear proteins are marked with numbers, whilst those proteins that are confined exclusively (or largely) to the cytoplasm are marked with letters and we refer to those from each stage with a subscript 1, 2 or 4, indicating that they are from the 1-, 2- or 4-cell stage. It is evident that there are marked differences in the distribution of newly synthesized proteins since there were several proteins at each stage that had been largely sequestered in the (pro)nuclei (Fig 2; lanesD and E: 1_1; 3₁, 5₁, 6₁, 7₁, 8], 10), 12] & 13, ; both lanes H: 1₂, 2₂, 4₂, 5₂, 12₂ & 13₂; both lanes K: 1₄, 3₄, 4₄, 12₄, 14₄ & 16₄), whilst other proteins were found predominantly in the cytoplasm and had been largely excluded from the (pro)nuclei (lanes B and C: for example A,, J,; lane G: for example A₂, H₂; lane J: for example A₄, J₄) during the 2−5 h period between labelling and nuclear isolation. There were also proteins that were found in both compartments.

There is no obvious difference in the proteins of the male and female pronuclei (Fig. 2; compare lanes D and E). The translation pattern changes quite dramatically during the 2-cell stage following embryonic gene activation (Flach et al. 1982) and this was paralleled by changes in the patterns of nuclear-associated proteins (Fig. 2, compare lanes D and E with H and K). The rate of protein synthesis is low during the 2-cell stage and thus the amount of incorporated label is lower than in the 1- or 4-cell embryo.

In the previous analysis of the localization of newly synthesized proteins in the early mouse embryo (Howe & Solter, 1979), no significant qualitative differences in the patterns of nuclear and cytoplasmic proteins were found. However, in that study, fractionation was achieved by centrifugation of cellular components. Comparing the previous study with the one reported here it appears that the isolation of early embryonic nuclei by micromanipulation is probably a more efficient way of obtaining relatively ‘clean’ nuclei that are free from cytoplasmic contamination. This, together with the fact that analysis was attempted by the use of two-dimensional gels, which make cross comparisons more difficult, may have contributed to the ambiguous results previously reported. By contrast, in the present study, we detected very marked differences in labelled proteins in isolated nuclei compared with cytoplasts (see Fig. 2). This difference would suggest that there was little, if any, cytoplasmic contamination of the karyo-plasts. Two further points about the analysis reported here are important: first, since we were dealing only with newly synthesized proteins, those that were already present in the cytoplasm or within nuclei (e.g. actin and histones) would not have been detected. Second, although one-dimensional gels make direct comparisons much easier, each band is almost certainly not a unique species but may well comprise several components. However, this one-dimensional analysis is a more convenient and reliable method for comparison as it indeed revealed marked differences in newly synthesized proteins that associated preferentially with nuclei and others that were largely excluded from nuclei. This observation clearly demonstrates selective sequestration of newly synthesized proteins in embryonic nuclei.

Since there was rapid accumulation and sequestration of proteins in early embryonic nuclei, it is possible that some of these proteins may be involved in the reprogramming of the hitherto relatively inert chromatin. Furthermore, it seems probable that zygotic proteins would associate with an introduced foreign nucleus, such as an 8-cell nucleus. Such an association possibly explains the observation that an 8-cell nucleus returned to 1-cell cytoplasm is initially transcriptionally silenced followed by activation of the hsp 68/70K genes (Howlett et al. 1987) which are normally the first embryonic gene products (Flach et al. 1982; Bensaude et al. 1983). From this present study, it appears that the hsp 68/70K proteins are distributed in both cytoplasmic and nuclear compartments (Fig. 2, lanes F−H, marked 7₂ and 8o), although at different phases of the cell cycle or under different conditions this distribution may be altered (Velazquez & Lindquist, 1984). It may be significant that qualitatively different translation products at each of the developmental stages analysed became localized in the nucleus concomitant with the transcriptional activation of the embryonic genome (Fig. 2, compare lanes D, E, H and K). It is possible that some of the proteins that were found in the nucleus were specifically modified forms of cytoplasmic proteins. For example, in the 1-cell embryo the nuclear proteins (Fig. 2, lanes D and E) marked 10, and 12, may be the modified forms of the cytoplasmic proteins G, and I, (Fig. 2, lanes A−C), respectively. In particular, I, is the interphase component of a 30K complex (Fig. 2, lanes A−C) that is phosphorylated during M-phase (Howlett, 1986), and therefore it is possible that the phosphorylated form (Fig. 2, lanes D and E, marked 12,) associates with chromatin. The heavily synthesized 35K protein (Fig. 2, lanes A−C, marked H,) that has been described previously (for example, see Howlett & Bolton, 1985) appeared to be protein marked 13, (Fig. 2, lanes D, E), with a molecular weight of about 22K, is particularly unstable (S.K.H. unpublished observations); this behaviour would be consistent with it being the ubiquiti-nated form of histone H2A (uH2A; Goldkopf et al. 1975), although we have no direct evidence to support this notion.

In view of the differential and complementary roles of the maternal and paternal chromosomes during mouse development (McGrath & Solter, 1984; Surani et al. 1984), it was interesting to determine if the two pronuclei accumulated and/or contained different proteins that may help explain their different developmental contributions. There were, however, no obvious differences detectable in the proteins sequestered within the male or female pronucleus, although this analysis only accurately reflects the more abundantly translated products. Furthermore, we have also failed to detect any differences in the protein content of parental pronuclei following iodination of total proteins (our unpublished observations). The reason for the variable number of ‘nucleoli’ within these early nuclei is unknown, but this is also a feature of amphibian oocyte GVs (Dreyer et al. 1983).

It would be interesting to compare the Xenopus GV-specific proteins with those of the mouse, by the use of specific antibodies. The fate of the nuclear proteins in the mouse during early mitoses to assess, for example, which proteins remain associated with chromosomes, remains to be determined. The differences in nuclear and cytoplasmic proteins reported in this study could help in the characterization and identification as well as the understanding of the functions of some of the translation products within the early mouse embryo.

We would like to thank Linda Notton for typing the manuscript. S.K.H. is in receipt of an MRC Training Fellowship.