Regulation of Protein Synthesis during Oogenesis, Oocyte Maturation and Early DevelopmentAs you know, protein synthesis is required for oocyte maturation and early development, but in many organisms initial protein synthesis utlilizes transcripts that are synthesized during oogenesis and stored for later utilization. Furthermore, the utilization of transcripts during development is very selective. What mechanisms select among transcripts and determines which ones will be translated during oogenesis, which will not, and when and where the stored transcripts will be utilized? This problem has been studied extensively in Xenopus, and we shall use this species to focus on this problem.
One mechanism controlling translatability of some oogenic messengers is the level of polyadenylation of those transcripts; the determinant of their polyadenylation is located in their 3' UTR. One class of messengers is translationally inert during oogenesis and is translationally activated during oocyte maturation (i.e., during the transition from prophase I to metaphase II of meiosis). Two cis-acting elements are necessary for this mode of regulation: the nuclear polyadenylation signal (AAUAAA; also called A2UA3) and a U-rich cytoplasmic polyadenylation element (CPE). The sequence of the CPE and its position relative to A2UA3 control the extent and timing of polyadenylation (Wormington, 1994; Vassalli and Stutz, 1995). Some transcripts respond to poly(A) tail length, whereas others require the act of polyadenylation itself.
Translation of the transcripts encoding the cell-cycle regulatory proteins cyclin and c-mos is controlled by polyadenylation; their translation - in turn - controls progression through the meiotic cell cycle. You will recall that maturation of the oocyte is triggered by progesterone, which is produced by the follicle cells surrounding the oocyte in response to pituitary gonadotropins. Multiple cyclins (A1, A2, B1, and B2) have been identified, which are synthesized during either oocyte maturation or after fertilization. Although the exact roles of the various cyclins are unknown, translation of cyclin B is necessary for the cleavage mitoses. Synthesis of c-mos is essential for maturation. (The functions of the cyclins and c-mos will be discussed futher by Dr. Lohka.) The synthesis of these various proteins is under very fine control. The messengers encoding these proteins are present in the full-grown oocyte, during maturation, and after fertilization, but there are dramatic differences in their translatability. Among the cyclins, cyclin A1 accumulation is very low in oocytes and increases during maturation, whereas both cyclin B1 and B2 accumulate to relatively high levels in oocytes (Kobayashi et al., 1991, Figs. 9 and 3 ). Cyclin A2, on the other hand, does not accumulate in significant amounts until after the mid-blastula transition (Howe et al., 1995, Fig. 4).
Sheets et al. (1994) examined the possibilty that translational activation of messengers encoding c-mos and cyclins A1, B1, and B2 during maturation is regulated by cytoplasmic polyadenylation. The lengths of poly(A) tracts on the endogenous transcripts were measured at different times after treatment of oocytes with progesterone. To monitor poly(A) length, total RNA was isolated and annealed to a DNA oligo complementary to a region 200 nucleotides upstream of the poly(A) tail. Half of each sample was also annealed to oligo(dT). Both samples were then treated with RNAse H, and the lengths of the 3' UTR were determined by Northern blotting, using a 3' UTR-specific probe. The length of the poly(A) was the difference in length between the RNAs treated with or without oligo(dT). Results are shown in Figs. 1 and 2 and summarized in Fig. 3 (Sheets et al., 1994). The data indicate a complex pattern of changes in poly(A) tail length; different transcripts lengthen to different extents with different kinetics: c-mos and cyclin B1 poly(A) increased in length by 70 and 220 nucleotides, respectively within 4 hours of progesterone treatement, whereas cyclins B2 and A1 increased by 80 and 100 nucleotides, respectively 5 hours later. Whereas the elongated poly(A) is retained on the cyclin messengers in 2-cell embryos, c-mos undergoes poly(A) shortening late in maturation and becomes completely deadenylated after fertilization.
Do the 3'UTRs of the transcripts possess the information that is sufficient to regulate polyadenylation? To examine this possibility, Sheets et al. microinjected RNAs containing only the terminal 100 nucleotides of these transcripts into oocytes and treated them with progesterone. The RNA was recovered and the length of poly(A) added to each was determined. The injected RNA fragments replicate the polyadenylation patterns of the endogenous transcripts, indicating that the 3' UTR is sufficient to regulate this behaviour (Sheets et al., 1994, Fig. 4).
Does the polyadenylation of these transcripts influence their translation? To examine this possibility, Sheets et al. fused the cyclin A1, B1, and c-mos UTRs to the luciferase coding region, synthesized RNA in vitro, injected the RNA into oocytes, and determined both the polyadenylation status and the level of luciferase activity in the presence and absence of progesterone (Sheets et al., 1994, Fig. 6). The ratio of luciferase activity in the matured vs. nonmatured oocytes was used as an index of translational stimulation. Mutations in the regions that control polyadenylation prevent translational activation. Sheets et al. concluded that each of the 3' UTRs stimulated translation during maturation as a result of their polyadenylation. In addition, the extent of translational stimulation can be correlated with the length of added poly(A). As shown in Figure 3 (Sheets et al., 1994), the chronology of poly(A) addition is quite different for some of these transcripts. Cyclins A1 and B1 represent two extremes; the former is polyadenylated several hours after the latter. Do the 3' UTRs exert this temporal control? The data in Fig. 7 (Sheets et al., 1994) suggest that they do.
The effects of poly(A) on translation are intriguing, because protein synthesis is initiated at the opposite, or 5', end of RNA. One hypothesis for the mechanism underlying this effect is that 3' polyadenylation induces 5' cap ribose methylation (Kuge and Richter, 1995). There is considerable evidence that not only poly(A) but also other 3' elements can exert profound influences on translation. It is clear that long-range interactions occur in RNA such that 3' elements can influence events occuring at the 5' end. One possibility is that the 3' end of RNA can influence the effectiveness of translation initiation factors in protein synthesis.
The messenger encoding the FGF receptor-1 (XFGFR-1) provides further evidence for long-range interactions in regulation of translation of oogenic mRNA. This receptor is involved in mesoderm induction, as demonstrated by mesoderm insufficiency and repression of mesoderm-specific gene expression when FGF signaling is inhibited in Xenopus embryos (Amaya et al., 1991, Cell 66: 257). FGF signaling is also required for activin-mediated induction of mesoderm (Cornell and Kimelman, 1994, Development 120: 453; LaBonne and Whitman, 1994, Development 120: 463). Thus, the appropriate blastomeres must display the appropriate surface receptors at the right time to be able to respond to the specific signals. Because mesoderm induction occurs before the onset of zygotic transcription, the FGF receptor must be encoded by oogenic messengers, and translation of those messengers must be regulated to ensure that the receptor is present in the right cells at the right time and in sufficient amounts.
The XFGFR-1 messenger is present in oocytes and early embryos. The messenger is not translated in immature oocytes; translation is initiated at oocyte maturation, coincident with a lengthening of the poly(A) tail (Fig. 1; Robbie et al., 1995). Robbie et al. examined the possibility that a cis-acting signal in the 3' UTR represses translation in the immature oocyte. Their results (Fig. 2) show that the 3' UTR is sufficient to repress translation and that the repression is specific to oocytes. To examine whether the inhibition by the 3' UTR is intrinsic to the UTR or whether an interaction between it and the coding sequence is necessary, chimeric constructs incorporating 3' UTR elements and a truncated human EGF receptor coding sequence (HED) were produced. As shown in Figure 3, the XFGFR-1 UTR was sufficient to inhibit translation.
Deletion analysis was conducted to determine which portion(s) of the UTR are responsible for translation inhibition. The results indicate that a 180 nucleotide proximal region was sufficient to inhibit translation. This portion of the UTR was coined the translation inhibitory element (TIE; Fig. 4, Robbie et al., 1995). Is polyadenylation involved in either translation inhibition or release of inhibition at oocyte maturation? To examine the possible involvement of polyadenylation in translation in immature oocytes, truncations to remove the synthetic poly(A) and the A2UA3 polyadenylation elements were made. As shown in Fig. 5 (Robbie et al., 1995), inhibition is not dependent upon downstream elements. RNA recovered from injected oocytes is the same length as the uninjected RNA, indicating that no polyadenylation had occurred.
Because the translation of endogenous XFGFR-1 mRNA is triggered at oocyte maturation, one would predict that the inhibitory effects of the TIE are reversed at maturation. Oocytes were injected with transcripts containing the TIE, treated with progesterone, and assayed for protein. As shown in Figure 6 (Robbie et al., 1995), the inhibition caused by TIE is relieved by progesterone treatment.
As shown in Figure 1 (Robbie et al., 1995), the endogenous RNA lengthened at oocyte maturation. Based upon the apparent effects of polyadenylation on recruitment of many kinds of RNA onto polysomes at oocyte maturation, it would be reasonable to assume that polyadenylation triggers the translation of XFGFR-1 in response to progesterone. However, as shown in Figure 6, the injected RNA showed no significant increase in size. (Keep in mind that no cytoplasmic polyadenylation elements are present in these transcripts. This result suggests that activation of XFGFR-1 translation at maturation does not require polyadenylation.
Once again, the question of how a 3' UTR element can regulate translation becomes relevant. Is repression of translation in oocytes dependent upon the binding of protein to TIE? A gel retardation assay was conducted to detect RNA-protein binding (Robbie et al., 1995, Fig. 7). Radiolabeled RNA was incubated with oocyte protein and subjected to electrophoresis. Proteins in the extract caused the RNA to shift; the shift pattern was distinct from that of the coding region. To examine the binding of protein to TIE, the authors conducted UV cross-linking of protein with radioactive TIE. The predominant labeled protein runs at 43kDa. Unlike the control probe, binding of the protein to TIE is competed out by excess cold probe, indicating the specificity of the binding of p43 to TIE. Further work will be necessary to determine how derepression of translation occurs. If p43 is responsible for repression, there is likely a mechanism triggered by progesterone that eliminates its repressive effects, probably by dissociating the TIE/p43 association. One common way to alter protein function is by changing its phosphorylation status. Because we know that a number of changes in protein phosphorylation occur at maturation, this is one possibility that should be examined.
See also overview of cytoplasmic polyadenylation by Joel Richter.
deMoor, C.H. and Richter, J.D. 1997. The Mos pathway regulates cytoplasmic
polyadenylation in Xenopus oocytes. Mol. Cell. Biol. 17: 6419-6426.