Differential accumulation of ribonucleotide reductase subunits in clam oocytes: the large subunit is stored as a polypeptide, the small subunit as untranslated mRNA.

Within minutes of fertilization of clam oocytes, translation of a set of maternal mRNAs is activated. One of the most abundant of these stored mRNAs encodes the small subunit of ribonucleotide reductase (Standart, N. M., S. J. Bray, E. L. George, T. Hunt, and J. V. Ruderman, 1985, J. Cell Biol., 100:1968-1976). Unfertilized oocytes do not contain any ribonucleotide reductase activity; such activity begins to appear shortly after fertilization. In virtually all organisms, this enzyme is composed of two dissimilar subunits with molecular masses of approximately 44 and 88 kD, both of which are required for activity. This paper reports the identification of the large subunit of clam ribonucleotide reductase isolated by dATP-Sepharose chromatography as a relatively abundant 86-kD polypeptide which is already present in oocytes, and whose level remains constant during early development. The enzyme activity of this large subunit was established in reconstitution assays using the small subunit isolated from embryos by virtue of its binding to the anti-tubulin antibody YL 1/2. Thus the two components of clam ribonucleotide reductase are differentially stored in the oocyte: the small subunit in the form of untranslated mRNA and the large subunit as protein. When fertilization triggers the activation of translation of the maternal mRNA, the newly synthesized small subunit combines with the preformed large subunit to generate active ribonucleotide reductase.

The small subunit of ribonucleotide reductase is encoded by one of the most abundant translationally regulated maternal RNAs in clam and sea urchin eggs.

In both clam oocytes and sea urchin eggs, fertilization triggers the synthesis of a set of proteins specified by stored maternal mRNAs. One of the most abundant of these (p41) has a molecular weight of 41,000. This paper describes the identification of p41 as the small subunit of ribonucleotide reductase, the enzyme that provides the precursors necessary for DNA synthesis. This identification is based mainly on the amino acid sequence deduced from cDNA clones corresponding to p41, which shows homology with a gene in Herpes Simplex virus that is thought to encode the small subunit of viral ribonucleotide reductase. Comparison with the B2 (small) subunit of Escherichia coli ribonucleotide reductase also shows striking homology in certain conserved regions of the molecule. However, our attention was originally drawn to protein p41 because it was specifically retained by an affinity column bearing the monoclonal antibody YL 1/2, which reacts with alpha-tubulin (Kilmartin, J. V., B. Wright, and C. Milstein, 1982, J. Cell Biol., 93:576-582). The finding that this antibody inhibits the activity of sea urchin embryo ribonucleotide reductase confirmed the identity of p41 as the small subunit. The unexpected binding of the small subunit of ribonucleotide reductase can be accounted for by its carboxy-terminal sequence, which matches the specificity requirements of YL 1/2 as determined by Wehland et al. (Wehland, J., H. C. Schroeder, and K. Weber, 1984, EMBO [Eur. Mol. Biol. Organ.] J., 3:1295-1300). Unlike the small subunit, there is no sign of synthesis of a corresponding large subunit of ribonucleotide reductase after fertilization. Since most enzymes of this type require two subunits for activity, we suspect that the unfertilized oocytes contain a stockpile of large subunits ready for combination with newly made small subunits. Thus, synthesis of the small subunit of ribonucleotide reductase represents a very clear example of the developmental regulation of enzyme activity by control of gene expression at the level of translation.

Translational regulation. Y the message is masked?

Much mRNA in Xenopus oocytes is translationally dormant for months until meiotic maturation and fertilization. This masking has been found to be coupled to transcription and mediated by the Y-box protein FRGY2.

Levels of free PABP are limited by newly polyadenylated mRNA in early Spisula embryogenesis.

The poly(A) tail of eukaryotic mRNAs regulates translation and RNA stability through an association with the poly(A)-binding protein (PABP). The role of PABP in selective polyadenylation/deadenylation and translational recruitment/repression of maternal mRNAs that occurs in early development is not fully understood. Here, we report studies including UV-crosslinking and immunoblotting assays to characterise PABP in the early developmental stages of the clam Spisula solidissima. A single, 70 kDa PABP, whose sequence is highly homologous to vertebrate, yeast and plant PABPs, is detected in oocytes. The levels of clam PABP are constant in early embryogenesis, although its ability to crosslink labelled poly(A) is 'masked' shortly after fertilisation and remains so until the larval stage. Full RNA-binding potential of PABP in embryo lysates was achieved by brief denaturation with guanidinium hydrochloride followed by dilution for binding and crosslinking or by controlled treatment of lysates with Ca(2+)-dependent micrococcal nuclease. Masking of PABP, which accompanies cytoplasmic polyadenylation in maturing oocytes and in in vitro activated oocyte lysates, is very likely due to an association with mRNAs that bear new PABP target binding sites and thus prevent protein binding to the labelled A-rich probe. Functional implications of these findings as well as the potential application of this unmasking method to other RNA-binding proteins is discussed.

Vg1 RBP intracellular distribution and evolutionarily conserved expression at multiple stages during development.

We have analyzed the expression and intracellular distribution, during oogenesis and embryogenesis, of Vg1 RBP, a protein implicated in the intracellular localization of Vg1 mRNA to the vegetal cortex of Xenopus oocytes. Vg1 RBP (protein) colocalizes with Vg1 RNA at all stages of oogenesis. Vg1 RBP RNA, however, localizes to the animal pole during late oogenesis, and remains in the animal blastomeres and ectodermal precursors until its zygotic transcription is activated, around stage 12. Vg1 RBP mRNA then becomes expressed throughout the neural epithelium. Vg1 RBP mRNA expression is also detected in what appears to be neural crest cells undergoing delamination and lateral migration. By tailbud stages, Vg1 RBP expression is present in the branchial arches, otic vesicle, pronephros, and along the neural tube. To examine the expression pattern in different species, we cloned the zebrafish homolog of Vg1 RBP by using a highly homologous EST clone to screen an embryonic cDNA library. In situ hybridization reveals that Vg1 RBP RNA localizes early in oogenesis to the animal pole. Although Vg1 RBP RNA is detected in all blastomeres of the early embryo, the expression pattern in the one day old zebrafish embryo is almost identical to that of the equivalent stage Xenopus embryo. These results indicate that the zygotic expression pattern is similar in frogs and fish, and that there is a conserved zygotic expression of Vg1 RBP distinct from its expression in the oocyte.

Identification of the stable free radical tyrosine residue in ribonucleotide reductase. A sequence comparison.

The small subunit of ribonucleoside diphosphate reductase contains a unique tyrosine radical and a binuclear iron center. An alignment of different primary structures of the small subunit in Escherichia coli, the marine mollusc Spisula solidissima, Epstein Barr and Herpes simplex viruses shows that regions comprising residues 115-122, 204-212 and 234-241 (in E.coli numbering) are strikingly similar and are likely to be recognized as functionally important. Two of 16 tyrosine residues and 2 of 8 histidine residues are conserved. We propose that Tyr-122 is responsible for radical stabilization and that His-118 and His-241 together with Glu-115 and Asp-237 or Glu-238 are ligands of the iron center.

Regulation of translation by specific protein/mRNA interactions.

This review will focus on cases of specific translational control by protein/RNA interactions in the 5'- or 3'-UTR of eukaryote mRNA where either the cis-acting RNA determinant or the trans-acting protein (or preferably both) have been identified with fair certainty. Examples of messages that are regulated by 5' motifs, which are proposed to occlude ribosome binding when bound by their specific factors, include ferritin and ribosomal protein mRNAs and the autoregulated thymidylate synthase and poly(A)-binding mRNAs. However, it has become increasingly evident recently that 3' UTR determinants and their specific binding proteins also regulate translation efficiency either directly, or indirectly via an influence on the polyadenylation status of the mRNA. It is still unclear how events at the 3' end of mRNA influence ribosome binding. Most, if not all, of the mRNAs known to be regulated by 3' UTR motifs are subject to regulation during early development or during differentiation such as several spermatocyte and oocyte mRNAs and erythroid lipoxygenase mRNA. To date, in all cases where translation is controlled directly by specific protein/mRNA interactions, the protein seems to act as a negative regulator, a translational repressor, whose binding to the specific site on the mRNA results in inhibition of initiation. The only cases of translational activation known so far concern internal initiation of translation of picornaviral RNAs, but this topic is beyond the scope of this review.

Control of translation of masked mRNAs in clam oocytes.

The pattern of protein synthesis changes soon after fertilization of clam oocytes. The most abundant of the mRNAs whose translation increases at this time encode ribonucleotide reductase and the A- and B-type cyclins. These mRNAs have been cloned and sequenced, yet their sequences do not show regions of similarity that could explain the masking mechanism. However, these mRNAs retain their 'masked' state in cell-free translation assays and their translation can be activated by gel filtration in high salt, which probably removes repressor proteins. A 'competitive unmasking' assay was used to identify the protein-binding regions of each mRNA. This involved adding short segments of antisense RNA that annealed to the mRNA and displaced the repressors. The unmasking regions in ribonucleotide reductase and cyclin A mRNAs revealed by this assay are 120-140 nt long and are located in the central portions of the 3' non-coding regions.

Regulated polyadenylation of clam maternal mRNAs in vitro.

During meiotic maturation of Spisula oocytes, maternal mRNAs undergo changes in translation and in the length of their poly(A) tails. In general, those mRNAs that are translationally activated, i.e., unmasked, become polyadenylated, while deactivated mRNAs lose their poly(A) tails. The activated class of mRNAs encode ribonucleotide reductase, cyclins A and B and histone H3, while the proteins that stop being made include tubulin and actin. Previously, we demonstrated that mRNA-specific unmasking can be brought about in vitro by preventing the interaction of protein(s) with central portions of the 3' noncoding regions (masking regions) of ribonucleotide reductase and cyclin A mRNAs. In this report, we show that clam egg extracts are capable of sequence-specific polyadenylation of added RNAs since the 3' untranslated regions (UTRs) of ribonucleotide reductase and histone H3 mRNAs are polyadenylated, while that of actin mRNA is not. In contrast, oocyte extracts, as in vivo, are essentially devoid of polyadenylation activity. We present an initial characterisation of the cis-acting sequences in the 3' UTR of ribonucleotide reductase mRNA required for polyadenylation. The results suggest that the sequences for cytoplasmic polyadenylation are more complex and extensive than those determined in vertebrates and that they may partly overlap with the masking regions.

Unmasking mRNA in clam oocytes: role of phosphorylation of a 3' UTR masking element-binding protein at fertilization.

During meiotic maturation or after fertilization of invertebrate and vertebrate oocytes, many of the quiescent stored mRNAs are recruited into polysomes. In the clam, Spisula solidissima, such masked messages include the abundant mRNAs encoding cyclin A and the small subunit of ribonucleotide reductase. We have previously shown that mRNA-specific unmasking of these two messages can be achieved in vitro, in oocyte cell-free extracts, by the addition of antisense RNAs corresponding to a fairly short (130-140 nucleotides) segment in their cognate 3' untranslated regions. We postulated that the antisense RNAs prevented the binding of a masking repressor protein (Standart et al., 1990). Here we report UV-crosslinking and gel retardation studies which show that the masking portions of the translationally regulated mRNAs bind an oocyte protein of 82 kDa (p82), which is phosphorylated after fertilization. This modification was accompanied by altered RNP complex formation in gel retardation assays. These changes presumably reflect the activation of translation of the masked mRNAs. The role of p82 phosphorylation in maternal mRNA unmasking was assessed in a novel in vitro activation system developed from clam oocytes, based upon the natural rise in pH which accompanies fertilization. Concomitant with mRNA unmasking, several kinases, including cdc2 and MAP kinases were activated in this system, as was p82 phosphorylation. Inhibitors of serine/threonine kinases, including 6-DMAP, staurosporine, and H7 inhibited p82 phosphorylation, whereas inhibitors of tyrosine kinases, protein kinase C, cAMP-dependent protein kinase, and p70s6k did not prevent this modification. A specific inhibitor of cdc2 kinase, p27Kip1, prevented p82 phosphorylation and translational activation, strongly suggesting that p82 modification is required for unmasking.

A conserved role of a DEAD box helicase in mRNA masking.

Clam p82 is a member of the cytoplasmic polyadenylation element-binding protein (CPEB) family of RNA-binding proteins and serves dual functions in regulating gene expression in early development. In the oocyte, p82/CPEB is a translational repressor, whereas in the activated egg, it acts as a polyadenylation factor. Coimmunoprecipitations were performed with p82 antibodies in clam oocyte and egg lysates to identify stage-regulated accessory factors. p47 coprecipitates with p82 from oocyte lysates in an RNA-dependent manner and is absent from egg lysate p92-bound material. Clam p47 is a member of the RCK/p54 family of DEAD box RNA helicases. Xp54, the Xenopus homolog, with bona fide helicase activity, is an abundant and integral component of stored mRNP in oocytes (Ladomery et al., 1997). In oocytes, clam p47 and p82/CPEB are found in large cytoplasmic mRNP complexes. Whereas the helicase level is constant during embryogenesis, in contrast to CPEB, clam p47 translocates to nuclei at the two-cell stage. To address the role of this class of helicase in masking, Xp54 was tethered via 3' UTR MS2-binding sites to firefly luciferase, following microinjection of fusion protein and nonadenylated reporter mRNAs into Xenopus oocytes. Tethered helicase repressed luciferase translation three- to fivefold and, strikingly, mutations in two helicase motifs (DEAD--> DQAD and HRIGR-->HRIGQ), activated translation three- to fourfold, relative to MS2. These data suggest that this helicase family represses translation of maternal mRNA in early development, and that its activity may be attenuated during meiotic maturation, prior to cytoplasmic polyadenylation.

Dual roles of p82, the clam CPEB homolog, in cytoplasmic polyadenylation and translational masking.

In the transcriptionally inert maturing oocyte and early embryo, control of gene expression is largely mediated by regulated changes in translational activity of maternal mRNAs. Some mRNAs are activated in response to poly(A) tail lengthening; in other cases activation results from de-repression of the inactive or masked mRNA. The 3' UTR cis-acting elements that direct these changes are defined, principally in Xenopus and mouse, and the study of their trans-acting binding factors is just beginning to shed light on the mechanism and regulation of cytoplasmic polyadenylation and translational masking. In the marine invertebrate, Spisula solidissima, the timing of activation of three abundant mRNAs (encoding cyclin A and B and the small subunit of ribonucleotide reductase, RR) in fertilized oocytes correlates with their cytoplasmic polyadenylation. However, in vitro, mRNA-specific unmasking occurs in the absence of polyadenylation. In Walker et al. (in this issue) we showed that p82, a protein defined as selectively binding the 3' UTR masking elements, is a homolog of Xenopus CPEB (cytoplasmic polyadenylation element binding protein). In functional studies reported here, the elements that support polyadenylation in clam egg lysates include multiple U-rich CPE-like motifs as well as the nuclear polyadenylation signal AAUAAA. This represents the first detailed analysis of invertebrate cis-acting cytoplasmic polyadenylation signals. Polyadenylation activity correlates with p82 binding in wild-type and CPE-mutant RR 3' UTR RNAs. Moreover, since anti-p82 antibodies specifically neutralize polyadenylation in egg lysates, we conclude that clam p82 is a functional homolog of Xenopus CPEB, and plays a positive role in polyadenylation. Anti-p82 antibodies also result in specific translational activation of masked mRNAs in oocyte lysates, lending support to our original model of clam p82 as a translational repressor. We propose therefore that clam p82/CPEB has dual functions in masking and cytoplasmic polyadenylation.

A 250-nucleotide UA-rich element in the 3' untranslated region of Xenopus laevis Vg1 mRNA represses translation both in vivo and in vitro.

Xenopus laevis Vgl mRNA undergoes both localization and translational control during oogenesis. Vg1 protein does not appear until late stage IV, after localization is complete. To determine whether Vg1 translation is regulated by cytoplasmic polyadenylation, the RACE-PAT method was used. Vg1 mRNA has a constant poly(A) tail throughout oogenesis, precluding a role for cytoplasmic polyadenylation. To identify cis-acting elements involved in Vg1 translational control, the Vg1 3' UTR was inserted downstream of the luciferase ORF and in vitro transcribed, adenylated mRNA injected into stage III or stage VI oocytes. The Vg1 3' UTR repressed luciferase translation in both stages. Deletion analysis of the Vg1 3' UTR revealed that a 250-nt UA-rich fragment, the Vg1 translational element or VTE, which lies 118 nt downstream of the Vg1 localization element, could repress translation as well as the full-length Vg1 3' UTR. Poly(A)-dependent translation is not necessary for repression as nonadenylated mRNAs are also repressed, but cap-dependent translation is required as introduction of the classical swine fever virus IRES upstream of the luciferase coding region prevents repression by the VTE. Repression by the Vg1 3' UTR has been reproduced in Xenopus oocyte in vitro translation extracts, which show a 10-25-fold synergy between the cap and poly(A) tail. A number of proteins UV crosslink to the VTE including FRGY2 and proteins of 36, 42, 45, and 60 kDa. The abundance of p42, p45, and p60 is strikingly higher in stages I-III than in later stages, consistent with a possible role for these proteins in Vg1 translational control.

Ca2+ is required for phosphorylation of clam p82/CPEB in vitro: implications for dual and independent roles of MAP and Cdc2 kinases.

During early development gene expression is controlled principally at the translational level. Oocytes of the surf clam Spisula solidissima contain large stockpiles of maternal mRNAs which are translationally dormant or masked until meiotic maturation. Fertilisation of the oocyte leads to rapid polysomal recruitment of the abundant cyclin and ribonucleotide reductase mRNAs at about the time they undergo cytoplasmic polyadenylation. Clam p82, a 3' UTR RNA-binding protein, and a member of the CPEB (cytoplasmic polyadenylation element binding protein) family, functions as a translational masking factor in oocytes and as a polyadenylation factor in fertilised eggs. In meiotically maturing clam oocytes, p82/CPEB is rapidly phosphorylated on multiple residues to a 92-kDa apparent size, prior to its degradation during the first cell cleavage. Here we examine the protein kinase(s) that phosphorylates clam p82/CPEB using a clam oocyte activation cell-free system that responds to elevated pH, mirroring the pH rise that accompanies fertilisation. We show that p82/CPEB phosphorylation requires Ca2+ (<100 microM) in addition to raised pH. Examination of the calcium dependency combined with the use of specific inhibitors implicates the combined and independent actions of cdc2 and MAP kinases in p82/CPEB phosphorylation. Calcium is necessary for both the activation and the maintenance of MAP kinase, whose activity is transient in vitro, as in vivo. While cdc2 kinase plays a role in the maintenance of MAP kinase activity, it is not required for the activation of MAP kinase. We propose a model of clam p82/CPEB phosphorylation in which MAP kinase initially phosphorylates clam p82/CPEB, at a minor subset of sites that does not alter its migration, and cdc2 kinase is necessary for the second wave of phosphorylation that results in the large mobility size shift of clam p82/CPEB. The possible roles of phosphorylation for the function and regulation of p82/CPEB are discussed.

Maternal mRNA from clam oocytes can be specifically unmasked in vitro by antisense RNA complementary to the 3'-untranslated region.

Clam oocytes display a striking difference in the pattern of protein synthesis on fertilization, which is maintained when cell-free extracts are assayed in the rabbit reticulocyte lysate. The mRNAs encoding ribonucleotide reductase and cyclin A (the major masked messages in the oocyte extracts) can be translationally activated by gel filtration in 0.5 M KCl, presumably by removal of repressor protein(s). When synthetic RNAs corresponding to different segments of ribonucleotide reductase and cyclin A mRNAs were added to oocyte extracts in 0.5 M KCl in a "competitive unmasking" assay, specific and complete unmasking of these mRNAs occurred on addition of antisense 3'-noncoding transcripts. The unmasking region in ribonucleotide reductase mRNA maps to a region of 134 nucleotides centered one-third of the way down the 3'-noncoding region.

Cyclin synthesis, modification and destruction during meiotic maturation of the starfish oocyte.

The pattern of protein synthesis in oocytes of starfish Marthasterias glacialis changes during 1-methyladenine-induced meiotic maturation. One of the newly synthesized proteins, a major 54-kDa polypeptide, was synthesized continuously after activation but was destroyed abruptly just before appearance of the polar bodies at each meiotic division. This protein thus resembles the cyclin proteins identified in cleaving sea urchin and clam embryos. RNA extracted from oocytes before and after maturation encoded virtually identical polypeptides when translated in the reticulocyte lysate. However, there was poor correspondence between the in vitro translation products and the labelling pattern of intact cells. There was no exact in vitro counterpart to the in vivo-labelled cyclin. Instead, a major polypeptide of 52 kDa was seen which appears to be a precursor of the 54-kDa form of cyclin. The 52-kDa polypeptide was identified as cyclin by hybrid arrest of translation. Cyclin mRNA is ot translated to a significant extent before oocyte activation and is present in oocytes as nonadenylated form. It becomes polyadenylated when the oocytes mature. This behavior is also seen in the case of the mRNA for the small subunit of ribonucleotide reductase, another abundant maternal mRNA whose translation is activated at maturation.

Autoregulation of poly(A)-binding protein synthesis in vitro.

The poly(A)-binding protein (PABP), in a complex with the 3'poly(A) tail of eukaryotic mRNAs, plays important roles in the control of translation and message stability. All known examples of PABP mRNAs contain an extensive A-rich sequence in their 5' untranslated regions. Studies in mammalian cells undergoing growth stimulation or terminal differentiation indicate that PABP expression is regulated at the translational level. Here we examine the hypothesis that synthesis of the PABP is autogenously controlled. We show that the endogenous inactive PABP mRNA in rabbit reticulocytes can be specifically stimulated by addition of low concentrations of poly(A) and that this stimulation is also observed with in vitro transcribed human PABP mRNA. By deleting the A-rich region from the leader of human PABP mRNA and adding it upstream of the initiator AUG in a reporter mRNA we show that the adenylate tract is sufficient and necessary for mRNA repression and poly(A)-mediated activation in the reticulocyte cell-free system. UV cross-linking experiments demonstrate that the leader adenylate tract binds PABP. Furthermore, addition of recombinant GST-PABP to the cell-free system represses translation of mRNAs containing the A-rich sequence in their 5'UTR, but has no effect on control mRNA. We thus conclude that in vitro PABP binding to the A-rich sequence in the 5' UTR of PABP mRNA represses its own synthesis.

Translation of 15-lipoxygenase mRNA is inhibited by a protein that binds to a repeated sequence in the 3' untranslated region.

During red blood cell differentiation, the mRNA encoding rabbit erythroid 15-lipoxygenase (LOX) is synthesized in the early stages of erythropoiesis, but is only activated for translation in peripheral reticulocytes. Erythroid LOX, which like other lipoxygenases catalyses the degradation of lipids, is unique in its ability to attack intact phospholipids and is the main factor responsible for the degradation of mitochondria during reticulocyte maturation. Strikingly, rabbit erythroid LOX mRNA has 10 tandem repeats of a slightly varied, pyrimidine-rich 19 nt motif in its 3'-untranslated region (3'-UTR). In this study we demonstrate, using gel retardation and UV-crosslinking assays, that this 3'-UTR segment specifically binds a 48 kDa reticulocyte protein. Furthermore, the interaction between the 3'-UTR LOX repeat motif and the 48 kDa protein, purified to homogeneity by specific RNA chromatography, is shown to be necessary and sufficient for specific translational repression of LOX as well as reporter mRNAs in vitro. To our knowledge this is the first case in which translation, presumably at the initiation step, is regulated by a defined protein-RNA interaction in the 3'-UTR.