Evidence for translation of HPRT enzyme on maternal mRNA in early mouse embryos

Studies on non-mammalian species have shown that messenger RNA synthesized during oogenesis (maternal mRNA) is utilized for protein synthesis after fertilization (see review by Davidson, 1976). In the sea urchin, translation on maternal mRNA continues through a number of cell divisions when transcription is inhibited by antinomycin D (Gross & Cousineau, 1964). In contrast to this situation the early development of mouse embryos is very sensitive to inhibitors of RNA synthesis (Monesi, Molinaro, Spalletta & Davoli, 1970; Golbus, Calarco & Epstein, 1973) suggesting that embryonic message is required. Indeed expression of the embryonic mouse genome within the first few cleavages may be observed at the biochemical and cellular levels (see reviews by McLaren, 1976; Johnson, 1981; Magnuson & Epstein, 1981). Clegg & Pikó (1982) have shown that large heterodisperse RNA is already synthesized at the 1-cell stage, and additionally some newly synthesized polyadenylated RNA is present in very low amounts and may therefore be mRNA (see also Levey, Stull & Brinster, 1978; Pikó & Clegg, 1982).

However, evidence has also been presented for stored maternal mRNA in mammalian eggs. Schultz (1975) showed that the amount of the poly A⁺ RNA fraction in rabbit eggs 10 h after fertilization is equal to that in non-fertilized eggs, and protein synthesis during this time is insensitive to almost total inhibition of RNA synthesis by a-amanitin. The RNA of unfertilized mouse ova also contains poly A tracts (Levey et al. 1978; Pikó & Clegg, 1982) and maternal RNA, including a poly A⁺ fraction, persists throughout mouse preimplantation stages (Bachvarova & DeLeon, 1980).

Braude, Pelham, Flach & Lobatto (1979) have demonstrated that certain proteins are translated on maternal message in the early mouse embryo. These workers see a group of proteins appearing at the early 2-cell stage after twodimensional electrophoresis. When mRNA synthesis is inhibited by α-amanitin the proteins still appear, even if eggs are fertilized in the presence of inhibitor (Flach et al. 1982). From in vitro translates it has been found that the mRNA coding for the proteins appearing at the 2-cell stage is already present in the unfertilized mouse egg, which may suggest that the message is‘masked’ in the egg (Flach et al. 1982; Cascio & Wassarman, 1982).

We have presented evidence previously for a maternally-derived increase in the activity of a specific enzyme, the X-coded HPRT (hypoxanthine phosphoribosyl transferase, E.C.2.4.2.8). The first suggestion that HPRT was maternally derived in early cleavage stages came from investigations into X-inactivation. At the morula stage enzyme activity is embryo coded, and since two X chromosomes are active in female (XX) embryos they have twice the HPRT activities of males (XY) (Monk & Harper, 1978; Epstein, Smith, Travis & Tucker, 1978; Kratzer & Gartler, 1978). Prior to this, from the 1-cell zygote to the early 8-cell stage, male and female embryos have equivalent HPRT activities (Monk & Kathuria, 1977) thus suggesting that the rise in enzyme activity that occurs over this period is maternally regulated (Monk & Harper, 1978; Monk, 1978). Confirmation of the maternal origin of HPRT came from the following findings. Monk & Harper (1978) showed that 8-cell embryos from XX mothers contain twice the activity of HPRT of those from XO mothers, reflecting a maternal effect of X-chromosome dosage during oogenesis (Epstein, 1972; Monk & Kathuria, 1977); and Burgoyne, Harper & Monk (unpublished) showed that HPRT activity increases between the 1-cell and the 4- to 8-cell stage even in YO embryos, i.e. in the absence of an X chromosome and thus an HPRT gene. The maternal regulation of this early increase in HPRT activity could occur either by the translation of maternally-derived mRNA coding for HPRT or by a change in the activity of presynthesized HPRT, possibly via post-translational modification of the enzyme.

This paper presents evidence that the maternally regulated increase in HPRT activity is due to translation of enzyme on maternal mRNA. The relationship between the HPRT-activity increase and both translation and transcription was studied at different times throughout preimplantation development. Protein synthesis was inhibited by culturing embryos in cycloheximide (Epstein & Smith, 1973); mRNA synthesis was inhibited by culture in a concentration of 10μg/ml of a-amanitin, which specifically blocks RNA polymerase II (Lindell et al. 1970; Levey & Brinster, 1978), and which is known to block the activation of the embryonic genome at the 2-cell stage (Flach et al. 1982).

Our results reveal some interesting points concerning the stability of HPRT enzyme and message. Although inhibitor studies suggest both are stable in early preimplantation stages, there is a significant and reproducible loss of HPRT activity at the time of transition from maternal- to embryo-coded enzyme. Degradation of several other maternally-inherited enzymes occurs at this time suggesting some common mechanism may be involved in the changeover from maternal to embryonic enzymes.

Mouse embryos were obtained from random-bred MFI females (Olac 1976 Ltd.). Female mice were superovulated by intraperitoneal injection of 5i.u. PMS (pregnant mare serum) followed 45–48 h later by 5i.u. HCG (human chorionic gonadotrophin), and were then caged with MFI males. Ovulation was assumed to occur 12h after the HCG injection (Gates & Beatty, 1954).

1-cell eggs were separated from cumulus cells by 10 to 15 min digestion at room temperature in hyaluronidase (300 i.u./ml) after which eggs were washed free of enzyme in PB1 medium (Whittingham & Wales, 1969). 2- to 8-cell embryos were flushed from the oviduct and morulae and blastocysts from the uterus with PB1. Eggs and embryos were either cultured (see below) or harvested directly. To harvest, zygotes were washed twice in PB1/PVP medium (PB1 with polyvinylpyrrolidone, 4 mg/ml, in place of albumin) and collected, singly or in groups, in 5 μ1 PB1/PVP in a 10 μ1 microcap, the ends of which were sealed by melting in a flame. Samples were stored at −70 °C. Supernatant extracts were prepared by freeze thawing three times followed by centrifugation at 4 °C.

Droplets of medium 16 (Whittingham, 1971), plus or minus inhibitors of protein or RNA synthesis, were prepared under paraffin oil in plastic culture dishes and equilibrated in a humidified 5 % CO2 in air atmosphere at 37 °C for at least an hour. Embryos from a number of litters were usually pooled and then added to culture droplets, via a wash droplet, in groups of between 5 and 20. Embryos were harvested after culture as above.

Solutions of cycloheximide (Sigma) in medium 16 (1, 10 or 100μg/ml) were prepared fresh before embryo culture. A stock solution of α-amanitin (1 mg/ml, Boehringer Mannheim) in medium 16 was aliquoted in 10, μl volumes and stored at –70 °C; prior to use 1ml medium 16 was added to the tube giving a final concentration of 10μg/ml.

HPRT was measured in embryo extracts as described in Monk & Harper, 1978. The specific activity of [³H]guanine sulphate was 700Ci/M unless otherwise stated. The incubation period was either 3 or 4 h at 37 °C. Male and female embryos around the morulae stage have different HPRT levels; however, embryos were assayed in pools of seven or more to give the mean value for both sexes combined. LDH activity was measured by the method of Brinster (1965).

Fig. 1 shows the profile for HPRT activity during preimplantation development, the enzyme increasing in activity during cleavage. The rise in HPRT activity is statistically significant even by the 2-cell stage (Table 1).

Fig. 2 shows that culture in the drug cycloheximide inhibits the rise in HPRT activity at all preimplantation stages from the 2-cell stage to blastocyst. The rise in enzyme activity is therefore dependent on protein synthesis, consistent with de novo synthesis of HPRT.

To test the effect of inhibiting mRNA synthesis, embryos were cultured in the presence of 10 μg/ml of a-amanitin. Fig. 3 compares the effect of α-amanitin treatment with cycloheximide treatment during a very early phase of development. Fertilized eggs (18 h post HCG) cultured for 24 h in α-amanitin show the same rise in HPRT activity as in the control, whereas the increase is inhibited by cycloheximide. As the significant rise in HPRT activity (t₁₀ = 19·956, P< 0·001) is insensitive to inhibition of transcription while being eradicated by inhibition of translation, the simplest conclusion is that this early increase in HPRT activity is due to translation of new enzyme on preexisting maternal message.

In order to test later stages of development, embryos were cultured in α-amanitin for 24 h periods at different (overlapping) stages as follows: 2-cell to 4/8-cell (40–64 or 46–70h post HCG); 4-cell to 8-cell (52–76 or 58–82h post HCG); 8-cell to morula and blastocysts (64–88 or 70–94h post HCG). HPRT activities of control and experimental cultures were compared after 12 and 24 h. The results are shown in Fig. 4.

Early 2-cell embryos of the MFI strain do not culture well in vitro. Nevertheless a significant increase in HPRT activity (t₄ = 3·33, P < 0·05) occurs over 24 h culture which is insensitive to α-amanitin (Fig. 4A), as it is from the 1-cell to the 2-cell stage shown in Fig. 3. However some caution should be exercised in interpreting this period of transition from maternal to embryonic mRNA. It is conceivable that some embryonic mRNA for HPRT is synthesized prior to commencement of culture of the 2-cell embryos in α-amanitin (Fig. 4A) and that the rise in HPRT activity occurs on both embryonic and maternal message.

From the late 2-cell stage onwards preimplantation embryos culture well. For these later stages of development, except the latest period (Fig. 4F), α-amanitin reduces the HPRT activity significantly after 24 h treatment, showing that newly synthesized embryonic mRNA is now required for the normal enzyme increase, α-amanitin in fact inhibits a change in rate of enzyme increase between 12 and 24 h after addition of the drug, indicating that the increased rate requires transcription. The delay in the effect of α-amanitin on the increase in enzyme activity suggests a period of at least 12 h between synthesis of new mRNA and its expression as active enzyme. In Fig. 4F, the latest period tested, there is no change in rate of enzyme activity increase. This may indicate cessation of further mRNA synthesis or enzyme degradation (see Fig. 2) as embryos enter the late blastocyst stage. In all these later stages (Fig. 4B, C, D, E, F) α-amanitin allows a rise in HPRT activity to continue throughout the 24 h of treatment, presumably by translation on stable mRNA present at the time of addition of the drug.

During the period of changeover from maternally- to embryonically-derived HPRT there is a transient loss in enzyme activity. Fig. 1 shows that after 64 h post HCG the HPRT activity rises steeply for a short time, dips at 72 h (when the embryonic population consists of compacted 8-cell and morulae stages), and then resumes the high rate of increase up to the blastocyst stage where it appears to plateau. The decrease at 72 h post HCG is reproducible but variable (see Figs 1 and 5 (HPRT)); HPRT activity may drop significantly (Fig. 5 (HPRT), t₂₈ = 2 · 574, P<0 · 02) or only slightly (Fig. 1) or it may plateau. A similar drop in activity is observed at the same time for adenine phosphoribosyltransferase (APRT, E.C.2.4.2.7; data not shown).

It was noted from the literature that a number of other enzymes that increase their activity during preimplantation development display a drop in activity (Fig. 5), the troughs coinciding at 72 h post HCG e.g. malate dehydrogenase, fructose 1,6-diphosphate aldolase (MDH, FDA; Epstein, Wegienka & Smith, 1969), aspartate aminotransferase (AA; Moore & Brinster, 1973), β -glucuronidase (Gus^h; Chapman, West & Adler, 1977). There is another class of enzymes that are maternally inherited at relatively higher activities. These enzymes are stable for days and then show an exponential decline in activity initiated at 72 h post HCG (see Fig. 5) e.g. lactic dehydrogenase, glucose 6-phosphate dehydrogenase (LDH, G6PD; Epstein et al. 1969), phosphoglycerate kinase (PGK; Kozak & Quinn, 1975). Indeed Spielmann, Erickson & Epstein (1974) have shown by immunological identification that the lowered activity of both LDH and G6PD is due to loss of enzyme molecules, and Epstein et al. (1969) have suggested that specific degradative processes may be involved. However HPRT enzyme is stable for 12 to 24 h in the presence of cycloheximide up to the late morula stage (Figs 2 and 3). Even transfer to the drug at 66 h post HCG and 70 h post HCG (by which time the decline is usually underway) gives no evidence for degradation (data not shown). This apparent contradiction would be resolved if cycloheximide itself inhibits the process of enzyme degradation leading to the loss of activity.

Lactic dehydrogenase (LDH) is inherited at high levels by the fertilized egg and shows striking degradation after 72 h post HCG (Epstein et al. 1969), the same time as the decrease in HPRT activity. As little or no synthesis of LDH occurs during preimplantation development we examined this enzyme as a model system to investigate the effect of cycloheximide on degradation. Fig. 6 shows that unlike the situation with HPRT, the drop in LDH activity is insensitive to cycloheximide, despite embryos being cultured in the drug for 20 h prior to the decline; by 80 h post HCG both control and experimental samples have the same lowered activity (t₈= 5·9942, P<0·01).

There are several lines of evidence from our laboratory that the increase in HPRT activity during early cleavage stages of the mouse is maternally derived (Monk & Kathuria, 1977; Monk & Harper, 1978; Monk, 1978; Burgoyne, Harper & Monk, unpublished). An activity increase is demonstrable in eggs from as early as 6 h postfertilization when there is very little synthesis of embryonic RNA (Clegg & Pikó, 1977; Pikó & Clegg, 1982; Young, Sweeney & Bedford, 1978). In addition activity increase occurs up to 64 h post HCG in the presence of α-amanitin but is inhibited by cycloheximide. The most likely and simplest interpretation of these results is that the number of active enzyme molecules increases due to translation of HPRT on stable maternal message up till the late 2-cell stage, and, as supported by X-chromosome dosage studies, probably until the 4- to 8-cell stage.

It is conceivable that the protein required to be synthesized is not HPRT, but some other protein that activates or modifies an HPRT precursor. The results would then suggest that this ‘activator’ is translated on maternal message.

The conclusion that the early increase in HPRT activity is due to translation on stable maternal mRNA is based on the assumption that α-amanitin is an active inhibitor of mRNA synthesis at this stage. Levey & Brinster (1978) have shown that 10μg/ml of α-amanitin inhibits polymerase II activity in the blastocyst within an hour, but no such precise information is available for other cleavage stages. As a labelled form of the drug is not commercially available the uptake of α-amanitin has not been tested directly. Nevertheless Braude et al. (1979) subjected fertilized eggs to a 6h in vitro pulse of α-amanitin treatment which resulted in retarded development after subsequent culture in control medium, and Flach et al. (1982) have shown that eggs fertilized in vitro in the presence of α-amanitin did not later synthesize new polypeptides characteristic of the late 2-cell stage.

Even in the presence of α-amanitin the activity of HPRT rises at all preimplantation stages (Figs 3 and 4), due to synthesis on stable pre-existing mRNA. If we assume there is less than a 6 h lag in effectiveness of the drug, and possibly less than 1 h, it appears that embryonic message is preformed a significant time before it is translated as there is an interval between exposure to α-amanitin and the inhibition of a new rate of increase. Our results suggest that embryonic message for HPRT is synthesized by the late 2-cell or early 4-cell stage. However, because there is a delay between transcription and translation, active HPRT enzyme is still synthesized on maternal mRNA until the 4- to 8-cell stage. The delay in translation of embryonic mRNA may be due to the time normally taken to process the message or be due to active ‘masking’ and ‘demasking’ of the message. Other workers have evidence of delays between transcription and translation of messages required at blastulation, the lag period ranging from a few hours (Braude, 1979) to over 24 h (Schindler & Sherman, 1981). The following implications for protein studies in early development are worth noting: (1) a gene may be actively transcribed long before translation of its product is identified, and (2) a gene may be switched off while increasing levels of its product are still being measured (important in studies of X-inactivation).

The present results, and those of Monk & Harper (1978), suggest that HPRT is translated on maternal mRNA up to the 4- to 8-cell stage, after which the message is embryonic. Around this proposed time of changeover there is a decrease in HPRT activity which consistently occurs at 72h post HCG. The coincidence in timing of the activity loss of HPRT (and APRT), and the seven other enzymes shown in Fig. 5, suggests some common mechanism may be responsible for removal of maternally-inherited enzymes. The decrease in HPRT activity was not observed in embryos incubated in cycloheximide. The reason for this is not known. In contrast the marked LDH decline is insensitive to cycloheximide. It is worth noting that the process of compaction does not appear to be required for enzyme activity decrease. Culture of early 8-cell embryos in medium without calcium or magnesium to inhibit compaction (Ducibella & Anderson, 1975) did not inhibit the drop in either HPRT or LDH activity (data not shown).

A consideration of the profile of the HPRT curve in Fig. 1 shows extrapolation of the known maternal and embryonic contributions to produce hypothetical curves (broken lines, Fig. 1) based on the assumption that maternal enzyme is removed. At 64 h post HCG nearly all HPRT enzyme is maternally derived and at 76 h post HCG all, or nearly all, HPRT enzyme is embryo coded (Monk & Harper, 1978). Between these two times, we have shown that HPRT enzyme is no longer translated on maternal mRNA and HPRT activity declines. The composite curve implies that maternally-derived HPRT and its message are specifically removed. With reference to the maternal message the results of Bachvarova & DeLeon (1980) support such a model. Following an initial postfertilization decrease, maternally-derived polyadenylated RNA (putative mRNA) remains stable throughout the second day of gestation in the mouse. After 60 h postfertilization (or 72 h post HCG) the amount of maternal mRNA again declines, roughly consistent with the time postulated in the composite curve. However this degradation may not be specific to maternal message. Non-specific degradation would also serve to ‘flush out’ maternal enzyme and maternal message as these have no means of being renewed.

Questions relating to this event are whether the embryo is capable of distinguishing its own gene products from those of its mother and, if so, whether the latter are removed specifically or not. In this respect preliminary experiments (Harper, unpublished) have revealed no differences in the Km values for the two substrates for HPRT made on maternal or embryonic mRNA. Irrespective of the mechanisms involved, it may be important for the embryo to control its differentiation exclusively from its own genome from the time of compaction onwards.

We wish to thank Peter Braude, Paul Burgoyne, Martin Johnson and Dallas Swallow for stimulating discussion, and Paul Burgoyne and Dallas Swallow for many helpful comments on the manuscript.