Tristetraprolin Inhibits Poly(A)-Tail Synthesis in Nuclear mRNA that Contains AU-Rich Elements by Interacting with Poly(A)-Binding Protein Nuclear 1

Background

Tristetraprolin binds mRNA AU-rich elements and thereby facilitates the destabilization of mature mRNA in the cytosol.

Methodology/Principal Findings

To understand how tristetraprolin mechanistically functions, we biopanned with a phage-display library for proteins that interact with tristetraprolin and retrieved, among others, a fragment of poly(A)-binding protein nuclear 1, which assists in the 3''-polyadenylation of mRNA by binding to immature poly(A) tails and thereby increases the activity of poly(A) polymerase, which is directly responsible for polyadenylation. The tristetraprolin/poly(A)-binding protein nuclear 1 interaction was characterized using tristetraprolin and poly(A)-binding protein nuclear 1 deletion mutants in pull-down and co-immunoprecipitation assays. Tristetraprolin interacted with the carboxyl-terminal region of poly(A)-binding protein nuclear 1 via its tandem zinc finger domain and another region. Although tristetraprolin and poly(A)-binding protein nuclear 1 are located in both the cytoplasm and the nucleus, they interacted in vivo in only the nucleus. In vitro, tristetraprolin bound both poly(A)-binding protein nuclear 1 and poly(A) polymerase and thereby inhibited polyadenylation of AU-rich element–containing mRNAs encoding tumor necrosis factor α, GM-CSF, and interleukin-10. A tandem zinc finger domain–deleted tristetraprolin mutant was a less effective inhibitor. Expression of a tristetraprolin mutant restricted to the nucleus resulted in downregulation of an AU-rich element–containing tumor necrosis factor α/luciferase mRNA construct.

Conclusion/Significance

In addition to its known cytosolic mRNA–degrading function, tristetraprolin inhibits poly(A) tail synthesis by interacting with poly(A)-binding protein nuclear 1 in the nucleus to regulate expression of AU-rich element–containing mRNA.     

The properties and function of rapidly-labelled nuclear RNA

Summary Nuclei were isolated from cultured cells of Acer pseudoplatanus L. previously pulse-labelled with [5-³H]uridine or [³²P]phosphate and the properties of the rapidly-labelled RNA were studied. Polyacrylamide gel electrophoresis showed ribosomal RNA precursors and processing intermediates with molecular weights of 3.4, 2.5, 1.4 and 1×10⁶ daltons, together with polydisperse RNA. The relative proportions of ribosomal RNA precursors and polydisperse RNA varied according to the length of the labelling period, but after 30 min approximately 90% of the radioactive RNA was polydisperse. The relationship between this polydisperse RNA and messenger RNA was investigated. The percentage of total nuclear RNA retained by chromatography on oligodeoxythymidylic acid-cellulose columns varied from 6% to 16% depending on the length of the labelling period. This RNA fraction, which has an adenylic acid content of approximately 45%, is assumed to represent RNA with polyadenylic acid sequences attached. A larger proportion of the nuclear polydisperse RNA lacked polyadenylic acid. Both types of polydisperse RNA were similar in size and during polyacrylamide gel electrophoresis migrated as broad peaks with an average molecular weight of approximately 10⁶ daltons. The polydisperse nuclear RNA that lacks polyadenylic acid was found to be similar in nucleotide composition to ribosomal RNA and is assumed to represent growing chains of ribosomal precursor RNA. After short labelling times the majority of the radioactivity incorporated into nuclear RNA is present in molecules of this type. This suggests that the designation of pulse-labelled polydisperse RNA as messenger RNA or precursor to messenger RNA solely on the basis of rapid labelling and size heterogeneity is unsound. The average molecular weight of the polyadenylic acid-containing messenger RNA from the cytoplasm was less than that of the corresponding nuclear RNA (6 and 9×10⁵ daltons respectively). This suggest either that the majority of the nuclear polyadenylic acid-containing RNA does not enter the cytoplasm, or if it does, that it first undergoes a reduction in size.Abbreviations rRNA ribosomal RNA - mRNA messenger - RNA poly(A), polyadenylic acid, poly(A) and poly(A) - RNA RNA with and without poly(A) sequences attached - poly(U) polyuridylic acid - oligo (dT)-cellulose cellulose with oligo deoxythymidylic acid covalently attached - C cytidylic acid - A adenylic acid - G guanylic acid - U uridylic acid     

Disappearance of stored polyadenylic acid and mRNA during early germination of radish (Raphanus sativus L.) embryo axes

Polyadenylic acid [poly (A)] is detected, characterized and quantitated in dry radish embryo axis RNA using a ³H poly (U) probe. The amount of poly (A) gradually decreases after the onset of soaking, and, after a few hours, recovers to the initial level. This variation is shown to result from the addition of two opposed phenomena: the decay of stored poly (A) and the accumulation of newly synthesized poly (A). Stored poly (A), as well as the in vivo protein synthesis coded for by preformed mRNA, decreases during early germination with a half-life of two hours. As a whole, these results demonstrate that at least a fraction of the stored mRNA is translated as soon as the seed is soaked and that its role is rapidly taken over by newly-made mRNA.Abbreviations Poly (A) (+) RNA polyadenylated RNA - Poly (A) polyadenylic acid - Poly (U) polyuridylic acid     

Effects of 3′deoxyadenosine and actinomycin D on RNA synthesis in toad bladder: Analysis of response to aldosterone

Summary Previous studies have shown that aldosterone increases transepithelial active Na⁺ transport after a latent period of about 60 min and incorporation of³H-uridine into polyadenylated RNA (poly(A)(+)RNA) (putatively poly(A)(+)mRNA) as early as 30 min after aldosterone addition. To assess the physiological importance of this pathway, the effects of 3deoxyadenosine and actinomycin D were compared in studies on the urinary bladder of the toadBufo marinus. 3deoxyadenosine (30 g/ml) only partially, though significantly, inhibited the aldosterone-dependent increase in Na⁺ transport measured as short-circuit current (scc). The incorporation of³H-uridine into poly(A) (+)RNA was inhibited by 70 to 80%. In contrast, Actinomycin D (2 g/ml) totally inhibited the aldosterone-dependent increase in scc, and the incorporation of³H-uridine into poly(A)(+)RNA by 68 to 75%. 3deoxyadenosine or actinomycin D alone had no significant effects on baseline scc, while inhibiting poly(A)(+)RNA to the same extent. The differential effects of deoxyadenosine and actinomycin on aldosterone-dependent Na⁺ transport may be related to their different sites of action on RNA synthesis: both drugs inhibited, to a similar extent, cytoplasmic poly(A)(+)mRNA; however, 3deoxyadenosine, in contrast to Actinomycin D, failed to inhibit poly(A)(-)RNA, sedimenting between 4S and 18S (putatively poly(A)(-)mRNA). We conclude that the mineralocorticoid action of aldosterone during the first three hours depends on the synthesis of both poly(A)(+)mRNA and poly(A)(-)mRNA.     

Size of the poly(A) region in mouse globin messenger RNA

A 9S RNA fraction from mouse reticulocytes, containing the active - and -globin mRNAs, has been isolated by hybridization of the polyadenylate regions in the mRNAs to oligo(dT)-cellulose. The adenylate-rich sequence isolated by limited RNase digestion of the globin mRNAs migrates between 4S and 5S RNA standards when co-electrophoresed on 12% polyacrylamide gels. Poly(A) standards, 28 and 84 nucleotides in length, showed anomolous migration relative to the 4S and 5S RNAs. The average size of the adenylate-rich sequence, estimated by its migration relative to the poly(A) standards, is about 50 nucleotides. The polyadenylate stretch in mouse globin mRNA is therefore much shorter than those found in other mRNAs.     

Poly(A) length,cytoplasmic adenylation and synthesis of poly(A)+ RNA in early mouse embryos

The poly(A) content of early mouse embryos fluctuates widely: after a transient increase in the one-cell embryo, there is a 70% drop in the two-cell and an approximately fivefold increase between the two-cell and early blastocyst stages (L. Pikó and K. B. Clegg, 1982, Dev. Biol.89, 362–378). To shed light on the significance of these changes, we analyzed the size distribution of total poly(A) from embryos at different stages of development by gel electrophoresis and hybridization with [³H]poly(U). The number-average size of poly(A) tracts varies only slightly, from 61 to 77 nucleotides, indicating that the changes in poly(A) content are due primarily to changes in the number of poly(A) sequences, i.e., the number of poly(A)+ mRNA. From these data, the number of poly(A)+ mRNA can be estimated as follows: ovulated egg, 1.7 × 10⁷; one-cell embryo, 2.4 × 10⁷; late two-cell, 0.7 × 10⁷; late eight-cell, 1.3 × 10⁷; and early blastocyst, 3.4 × 10⁷. These results suggest the elimination of the bulk of maternal poly(A)+ mRNA at the two-cell stage, to be replaced by newly synthesized mRNA derived from the embryonic genome. To study the synthesis of poly(A)+ mRNA, we cultured mouse embryos in vitro with [³H]adenosine and analyzed the labeled poly(A)+ RNA as to molecular size, length of the poly(A) tail, and relative distribution of label in poly(A) vs internal locations. We observed an active incorporation of label into large-molecular-weight (average size about 2 kb) poly(A)+ RNA at all stages from the one-cell to the blastocyst. However, in the one-cell embryo, about 70% of the label was localized in the poly(A) tail, suggesting cytoplasmic polyadenylation, and only about 30% was localized in the remainder of the molecule, suggesting the complete new synthesis of a small amount of poly(A)+ RNA. Differences in the size distribution of the labeled poly(A) as compared with the total poly(A) in the one-cell embryo indicate that the labeling is not due to a general turnover of poly(A) tails, but rather to the polyadenylation of previously nonpolyadenylated, stored RNA. Significant new synthesis of poly(A)+ RNA is evident from the two-cell stage onward and most likely accounts for the sharp rise in the number of poly(A)+ RNA molecules by the early blastocyst stage.     

Terminal completion of poly(A) synthesis in sea urchin embryos.

The kinetics of accumulation of nuclear and cytoplasmic poly(A) have been determined in sea urchin blastulas and gastrulas, stages when essentially all mRNA is synthesized de novo in the nucleus. A majority of the labeled poly(A) is found in the cytoplasmic fraction after a brief pulse. The ratio of radioactive AMP to adenosine in pulse-labeled nuclear, cytoplasmic, and polyribosomal poly(A) is considerably less than the number average length of the labeled poly(A), indicating that there is 3′-terminal addition of adenosine to previously synthesized poly(A). The size distribution of pulse-labeled, terminally elongated poly(A) in the cytoplasm is similar to that of the largest nuclear poly(A) rather than the steady-state size distribution of cytoplasmic poly(A), which is smaller and more heterogeneous. The most likely interpretation of these results is that there is a predominant 3′ terminal addition of short tracts of adenosine to poly(A) attached to nuclear RNA just before or during entrance of this RNA into the cytoplasm. In this respect, much of the 3′ terminal addition may be thought of as terminal completion of poly(A) synthesis.     

In-vitro translation of messenger-RNA from developing bean leaves. Evidence for the existence of stored messenger-RNA and its light-induced mobilisation into polyribosomes

Poly(A)-containing messenger RNA was purified from polyribosomes isolated from the primary leaves of 7-day-old dark-grown seedlings of Phaseolus vulgaris var. Masterpiece. Analysis of the messenger RNA on 2.4% polyacrylamide gels showed that it consists of a heterogeneous population of molecules with an average molecular weight of 500,000. The nucleotide composition of the RNA was 16.0% cytidylic acid, 39.4% adenylic acid, 21.3% guanylic acid and 23.2% uridylic acid. Based on the degree of resistance of the RNA to digestion with ribonucleases A and T₁ the average length of the poly(A) sequence was calculated to be 120 nucleotides. No significant differences in mobility in polyacrylamide gels, nucleotide composition or polyadenylic acid content were found between the poly(A)-containing mRNA from polyribosomes of primary leaves of dark-grown plants and those given a 16 h white light treatment. Purified poly(A)-containing mRNA was shown to direct the incorporation of [³⁵S]methionine into proteins in an in vitro protein-synthesising system from wheat germ. The protein products were fractionated according to molecular size by electrophoresis in 15% polyacrylamide/urea/SDS gels and the protein bands were detected by fluorography. Messenger RNAs directing the synthesis of three polypeptides with molecular weights of 34,000, 32,000 and 25,000 were detected in polyribosomes of plants following white light treatment. These messenger RNAs were absent, or present in much lower amounts, in polyribosomal messenger RNA from leaves of dark-grown plants, although they were present in total cell poly(A)-containing RNA. This indicates that certain messenger RNAs may be stored in the dark and that light stimulates these RNAs to engage in polyribosome formation. Continuous far-red (730 nm) irradiation for 4 h also caused the appearance of these messenger RNAs in the polyribosomes although 5 min red light followed by 4 h darkness had little effect. This suggests that phytochrome acting in the high energy mode, may be the photoreceptor responsible for initiating the response.Abbreviations mRNA messenger-RNA - rRNA ribosomal RNA - oligo (dT) oligo (deoxythymidylic acid) - poly(A) polyadenylic acid - EDTA ethylenediamine-tetra-acetic acid - HEPES N-2-hydroxyethylpiperazine-N-2-ethane-sulphonic acid - SDS sodium dodecyl sulphate     

Cytoplasmic poly(ADP-ribose) polymerase and poly(ADP-ribose) glycohydrolase in AEV-transformed chicken erythroblasts

Poly(ADP-ribose) polymerase and poly(ADP-ribose) glycohydrolase activities were both investigated in chicken erythroblasts transformed by Avian Erythroblastosis Virus. Respectively 21% and 58% of these activities were found to be present in the post-mitochondrial supernatant (PMS). Fractionation of the PMS on sucrose gradients and poly(A⁺) mRNA detection by hybridization to [³H] poly(U) show that cytoplasmic poly(ADP-ribose) polymerase is exclusively localized in free mRNP. The glycohydrolase activity sedimented mostly in the 6 S region but 1/3 of the activity was in the free mRNP zone. Seven poly(ADP-ribose) protein acceptors were identified in the PMS in the Mr 21000–120000 range. The Mr 120000 protein corresponds to automodified poly(ADP-ribose) polymerase. A Mr 21000 protein acceptor is abundant in PMS and a Mr 34000 is exclusively associated with ribosomes and ribosomal subunits. The existence of both poly(ADP-ribose) polymerase and glycohydrolase activities in free mRNP argues in favour of a role of poly(ADP-ribosylation) in mRNP metabolism. A possible involvement of this post translational modification in the mechanisms of repression-derepression of mRNA is discussed.Abbreviations ADP-ribose adenosine (5) diphospho(5)--D ribose - poly(ADP-ribose) polymer of ADP-ribose - mRNP messenger ribonucleoprotein particles - PMSF phenylmethylsulfonyl fluoride - LDS lithium dodecyl sulfate - TCA trichloroacetic acid     

Free cytoplasmic messenger ribonucleoprotein complexes from rabbit reticulocytes

Free cytoplasmic globin mRNA containing mRNP-particles were isolated from rabbit reticulocytes by zonal sucrose gradient centrifugation and their properties were compared with mRNP particles isolated in the same way from EDTA-dissociated reticulocyte polyribosomes. The average poly(A)-length of 9S mRNA from free cytoplasmic mRNP was 17–20 nucleotides being about two times shorter than the average poly(A)-length of polysomal 9S mRNA. The protein composition of the free cytoplasmic mRNP particles disclosed the absence of the 76,000 dalton protein which is associated with the 3poly(A)-segment of polysomal globin mRNA. It was concluded that free cytoplasmic mRNP-particles from rabbit reticulocytes can be classified as old mRNP in a post-translational phase. Free cytoplasmic mRNPs were translated in heterologous cell-free systems as well as in Xenopus laevis oocytes. Addition of hemin stimulated the synthesis of -globin in all systems, while the presence of the cap analogue m⁷G(5)p inhibited translation of free cytoplasmic mRNA completely. The latter finding suggested that free cytoplasmic mRNA has a 5 terminal cap. Shortening of the poly(A)-segment with concomitant loss of the 76,000 dalton protein may lead to less efficient translation of free cytoplasmic mRNP.     

Minimal methylation of yeast messenger RNA

Most, if not all, yeast mRNAs are capped at their 5-terminus by m⁷G. Apart from m⁷G no other methylated nucleotides could be detected in poly (A)⁺ mRNA isolated from yeast polysomes.Abbreviations used poly (A)⁺ mRNA messenger RNA containing poly (A) - poly (A)^- RNA RNA lacking poly (A) - m⁷G N₇-methyl guanosine - N_m any 2-0 methylated nucleoside - mN any basemethylated nucleoside     

Why,when and how does the poly(A) tail shorten during mRNA translation?

• 1.1. The length of the poly(A) tail at the 3'-end of mRNA may control protein synthesis by bringing the 3'-end in close proximity to the 5'-end of the noncoding region as well as increasing the duration of mRNA translation by its binding to the poly(A) binding protein.
• 2.2. The rate-limiting step in the decay of the body of the message is the shortening of a long poly(A) tail during mRNA translation. The shortening of the poly(A) tail occurs during pre-elongation in the protein synthesis cycle.
• 3.3. The shortening of the poly(A) tail during mRNA translation may not involve RNase activity, however poly(A) binding protein seems to play a role, at least in part, in shortening of the poly(A) tail.

     

Characterization of two SV40 early mRNAs and evidence for a nuclear "prespliced" RNA species.

Using in vitro translation of sucrose-gradient fractionated cytoplasmic mRNA from SV40-infected cells, we have shown that a deletion in the region mapping between 0.54--0.59 reduced the size of mRNA for small-t but not the size of mRNA for large-T. Mutants with a deletion in this region were shown to produce in vivo either shortened small-t or no small-t, and normal large-T. Similarly, in vitro translation of poly(A)+cytoplasmic RNA from cells infected with these mutants gave the same results. On the other hand in vitro translation of poly(A)+nuclear RNA from the mutants which made no small-t produced a small-t derivative possibly synthesized from a prespliced RNA species. We have also shown that poly(A)+nuclear RNA from mutant dl 2122 produced two small-t related proteins: one of these (MW: 11K) probably represents the product of a "prespliced" RNA, the other (MW: 17K) which is also found in the cytoplasm represents the product of the mutant specific small-t mRNA.