Defining a novel cis-element in the 3'-untranslated region of mammalian ribonucleotide reductase component R2 mRNA. cis-trans-interactions and message stability.

Mammalian ribonucleotide reductase is a highly regulated activity essential for DNA synthesis and repair. The 3′-untranslated region (3′-UTR) of mammalian ribonucleotide reductase R2 mRNA has been implicated in the tumor promoter, 12-O-tetradecanoylphorbol-13-acetate-mediated stabilization of mouse BALB/c 3T3 R2 message. We investigated the possibility that the 3′-UTR contains regulatory information for R2 mRNA turnover. Using 3′-end-labeled RNA in gel shift and UV cross-linking analyses, we detected in the 3′-UTR a novel 9-nucleotide cis-element, 5′-UCGUGUGCU-3′, which interacted with a widely distributed cellular cytosolic protease-sensitive factor(s) in a sequence-specific manner to form a 45-kDa R2 binding protein complex. The binding activity was redox-sensitive and down-regulated by 12-O-tetradecanoylphorbol-13-acetate and okadaic acid in a dose-dependent manner. Insertion of a 154-base pair fragment containing the cis-element led to markedly reduced accumulation of chloramphenicol acetyltransferase hybrid mRNA relative to the same insert carrying a series of G → A mutations within this element that eliminated binding. We suggest that the 9-nucleotide region functions as a destabilizing element. These results provide a model for ribonucleotide reductase gene expression through a novel and specific mRNA cis-trans-interaction involving a phosphorylation signal pathway that leads to changes in the stability of R2 message.

Mammalian ribonucleotide reductase is a highly regulated activity essential for DNA synthesis and repair. The 3-untranslated region (3-UTR) of mammalian ribonucleotide reductase R2 mRNA has been implicated in the tumor promoter, 12-O-tetradecanoylphorbol-13-acetate-mediated stabilization of mouse BALB/c 3T3 R2 message. We investigated the possibility that the 3-UTR contains regulatory information for R2 mRNA turnover. Using 3-end-labeled RNA in gel shift and UV cross-linking analyses, we detected in the 3-UTR a novel 9-nucleotide cis-element, 5-UCGUGUGCU-3, which interacted with a widely distributed cellular cytosolic proteasesensitive factor(s) in a sequence-specific manner to form a 45-kDa R2 binding protein complex. The binding activity was redox-sensitive and down-regulated by 12-Otetradecanoylphorbol-13-acetate and okadaic acid in a dose-dependent manner. Insertion of a 154-base pair fragment containing the cis-element led to markedly reduced accumulation of chloramphenicol acetyltransferase hybrid mRNA relative to the same insert carrying a series of G 3 A mutations within this element that eliminated binding. We suggest that the 9-nucleotide region functions as a destabilizing element. These results provide a model for ribonucleotide reductase gene expression through a novel and specific mRNA cistrans-interaction involving a phosphorylation signal pathway that leads to changes in the stability of R2 message.
Mammalian ribonucleotide reductase is composed of two dissimilar dimeric protein components, proteins R1 and R2, which are required to catalyze the direct reduction of ribonucleoside diphosphates to the corresponding deoxyribonucleotides, a rate-limiting step in the synthesis and repair of DNA (1,2). The importance of ribonucleotide reductase for cell proliferation is further emphasized by the observation that the mechanisms controlling ribonucleotide reductase gene expression may be altered in some malignant conditions (3). Ribonucleotide reductase may play an important role in the critical early events involved in the process of tumor promotion by TPA 1 (4) and appears to be important in an aberrant growth factor signal pathway in Ha-ras-transformed cells (5). Clearly, obtaining more information about the mechanisms that control ribonucleotide reductase gene expression is important for understanding DNA synthesis, DNA repair, and cell proliferation (2,6).
The regulation of mRNA stability has emerged as an important control mechanism of gene expression. Although the mechanisms that alter mRNA stability of different genes have unique features, it appears that in each case specific RNA sequences, mainly located in the 3Ј-untranslated regions (3Ј-UTRs) of mRNAs, are required for the interactions with specific protein factors (7)(8)(9)(10)(11)(12)(13). For example, ribonucleotide reductase R2 mRNA stability is elevated by TPA (8). Although important details concerning the mode of action of TPA are still unknown, the prevailing hypothesis is that TPA exerts its diverse effects through the activation of protein kinase C, a high affinity target for TPA (14 -16). Although it is now evident that TPA can affect gene expression posttranscriptionally through altering message stability (8,9,17,18), relatively little is known about the molecular mechanisms of the cis-elements and transfactors involved. We have recently shown that a 20-nt fragment (FokI-HgiA) in the 3Ј-UTR of R2 mRNA is involved in the TPA-mediated stabilization of R2 message (8). The present study was intended to define the unique TPA-responsive ciselement in the 3Ј-UTR that leads to alterations in R2 mRNA stability.

EXPERIMENTAL PROCEDURES
Cell Culture-Mouse BALB/c 3T3, 10T 1 ⁄2, COS7, C1 and C2 fibrosarcoma, human HeLa S3, and cells transfected with chloramphenicol acetyltransferase (CAT) R2 3Ј-UTR hybrid plasmids were routinely cultured in minimal essential medium supplemented with antibiotics and 10% (v/v) fetal bovine serum (19). For determining CAT/R2 3Ј-UTR hybrid mRNA levels and the cis-element binding activity, exponentially growing cells were treated for different times with various concentrations of TPA (Sigma), okadaic acid (Upstate Biotechnology Inc., Lake Placid, NY), and dimethyl sulfoxide (Me 2 SO). Me 2 SO was used to dissolve TPA and OKA; control cells received medium containing 0.01% Me 2 SO alone. Cells were also treated with 10 g/ml actinomycin D (Sigma) to block transcription and were harvested as described previously (20).
Preparation of Protein Extracts from the Cytosol and the Nucleus-Cells transferred to Eppendorf tubes were briefly centrifuged for 1 min, resuspended in hypotonic buffer (25 mM Tris-HCl, pH 7.9, 0.5 mM EDTA), and lysed by repetitive cycles of freeze-thaw. Nuclei and cytoplasmic extracts were obtained as described previously (8), and protein concentrations were determined by the Bio-Rad protein assay kit as described by Bio-Rad. 2 In Vitro Transcription-A summary of the construction of the in vitro transcription plasmid p3Ј-UTR containing the full-length 3Ј-UTR of R2 cDNA from mouse TA3 cells has been previously described (8). In brief, run-off RNA transcripts were produced by T7 polymerase activity (Boehringer Mannheim) from 1 g of digested cDNA plasmids as described previously (21). RNA transcripts were produced and extracted to a specific activity of approximately 3 ϫ 10 8 cpm/g of RNA (10).
3Ј-End Labeling of Synthetic RNA Oligonucleotides-The RNA oligoribonucleotides purchased from Dalton Chemical Laboratories Inc. (Toronto, Canada) were synthesized on the Applied Biosystems (ABI) model 392 DNA/RNA synthesizer using phosphoramidite RNA monomers according to the manufacturer's instructions. 3 The ribonucleotides were labeled at their 3Ј-ends by a previously described method (22). The 3Ј-end-labeled RNA oligoribonucleotides were gel-purified, extracted, reconstituted in RNase-free water (10), and stored at Ϫ70 C.
RNA Mobility Shift Assay-Binding reactions were performed as described previously (21) with 10 -40 g of cytosolic and nuclear protein extract and 50 ϫ 10 3 cpm of 32 P-labeled RNA transcript or oligoribonucleotides. However, 50 units of RNase T1 (Boehringer Mannheim) were added when the R2 3Ј-UTR was used (but not for binding reactions with the 3Ј-end-labeled oligoribonucleotides). In some assays cytoplasmic extracts were preincubated with compounds such as diamide and N-ethylmaleimide for 10 min before the addition of the riboprobe.
UV Cross-linking Analyses-RNA-protein binding reactions were carried out as described above. Following the addition of heparin, samples were placed in microtiter wells placed on ice and UV-cross-linked for 10 min in a Stratalinker Chamber (Stratagene), at a 250-mJ energy level. The samples were separated by electrophoresis on a 10% SDSpolyacrylamide gel (21). The RNA-protein complexes may dissociate to some degree during electrophoresis, resulting in a diffuse banding pattern (23).
Northern Blot Analysis and Half-life Measurements-Total cellular RNA was extracted from drug-treated and untreated mouse BALB/c 3T3 transfectants by a rapid method of RNA preparation (24). For half-life measurements, the cells were treated with 10 g/ml actinomycin D. RNA was isolated at different times and analyzed for CAT and CAT/R2 3Ј-UTR hybrid transcripts. Twenty g of total cellular RNA was electrophoresed through 1% formamide-agarose gels and blotted onto Nytran nylon membranes. The blots were prehybridized and probed with the 32 P-labeled HindIII-BanI fragment of CAT cDNA, derived from the plasmid pSVlacOCAT (8). Determination of sample loading was performed by probing with glyceraldehyde-3-phosphate dehydrogenase cDNA (5). The mRNA half-life was estimated as described previously (7,10).
Plasmid DNA Constructs and Transfection-The plasmid pCAT and the pCAT/R2 3Ј-UTR hybrid construct were similarly derived as described previously (8), with the notable exception that in this study the expression vector pECE polyadenylation signal fragment (25) was not deleted. This confers an increased half-life upon the pCAT transcript. Transfection of mouse BALB/c 3T3 cells was carried out by the calcium phosphate method as described previously (26), using the plasmid pSV2neo as a selectable marker. Northern blot analysis (7, 10) and CAT assays (27) were performed to screen for permanently transfected cell lines.
The PCR (40 cycles) was performed in a 100-l total volume, and the reacting mixtures contained 50 ng of DNA template, 40 pmol of DNA primers, and 5 units of Taq DNA polymerase. The samples were gelpurified and subsequently digested with the restriction endonucleases, SalI/XbaI. Digests were heat-inactivated at 65°C for 15 min and ligated into the pCAT plasmid to generate pCAT/R2 normal and pCAT/R2 mutant. The nucleotide sequence of these fragments were confirmed by DNA sequencing (Sequenase). The wild-type DNA fragment (R2 normal) containing 154 base pairs of the 3Ј-UTR of R2 mRNA (nt positions 1224 -1378), with the putative R2 binding protein complex (R2BP) cis-element, was generated by PCR with the plasmid pCD10 (8), harboring the R2 cDNA and the overlap extension primers, 5Ј-ACA-CACGTCGACGCTGACTTCTAAGTAACTC-3Ј and 5Ј-CGACACATC-GATGCAGTCAGCACAGATCTACACAC-3Ј. Similarly, oligonucleotides, 5Ј-ACACACGTCGACGCTGACTTCTAAGTAACTGATCATAT-ACTCTTCG-3Ј and 5Ј-CGACACATCGATGGAGTCAGGACAGATC-TACACAC-3Ј were used as primers to generate the 154-base pair DNA fragment of the R2 3Ј-UTR containing the mutated cis-element (R2 mutant). The underlined sequences shown above represent the SalI and XbaI linker sequences.

Identification of a Novel cis-Element in the 3Ј-UTR of Ribo-
nucleotide Reductase R2 mRNA-To locate the binding site sequence within the 20-nt (FokI-HgiA) fragment of the 3Ј-UTR of R2 mRNA that forms the 45-kDa R2BP, we generated a series of synthetic oligoribonucleotides corresponding to different segments of the 20 nt. These RNA fragments were 3Ј-endlabeled and used in standard gel shift assays. The data summarized in Fig. 1, indicate that the 20-nt fragment forms 3 Applied Biosystems User Bulletin 47. similar RNA-protein complex shifts as the full-length 3Ј-UTR of R2 mRNA (Fig. 1C). The RNA fragment, 5Ј-UCGUGUGCU-3Ј, is identified as the only binding site within the 20-nt fragment, shown by the formation of the RNA-protein complex, when the corresponding riboprobe was incubated with cytoplasmic extracts from BALB/c 3T3 fibroblasts (Fig. 1D, lane 6). This putative cis-element appears to be unique, since it does not resemble any known, previously described RNA cis-element, such as the AU-rich pentamer (11), the iron response element (28), the amyloid precursor protein mRNA 29-nt stability element (13), or the ribonucleotide reductase R1 mRNA cis-element (10). Furthermore, to determine the minimal ciselement sequence, formation of the R2BP was not observed for shorter oligoriboprobes under the same conditions that lead to binding activity with the entire 9-nt cis-element (Fig. 1E). Although it is possible that under other experimental conditions, such as higher RNA concentrations, these partial sequences may exhibit some binding activity, these results do suggest that the entire 9-nt sequence is optimal for binding.
Characterization and Sequence Specificity of the cis-Element Binding Activity in the R2 3Ј-UTR-To determine if the binding factor was a protein, cytosolic extract from mouse BALB/c 3T3 cells was incubated with the 3Ј-end-labeled putative cis-element fragment in the presence of proteinase K (40 units/ml). This completely abolished R2BP formation (Fig. 2A, lane 1). The cis-element binding activity was not detected in nuclear extracts but was observed in cytosolic extracts ( Fig. 2A, lanes 2  and 3). These results indicate that the complex detected in these experiments is composed of RNA and polypeptides and that the mechanism involved in the regulation of R2 mRNA stability is likely a postnuclear event. Using the 20-nt and full-length R2 3Ј-UTR RNA transcript as specific competitors in gel shift assays with the cis-element sequence as a riboprobe led to inhibition of the formation of the R2BP (Fig. 2B, lanes 1  and 2). The addition of 100 M excess of the human c-myc mRNA 3Ј-UTR fragment (12) did not prevent the RNA-protein complex formation (Fig. 2B, lane 3). Since the c-myc mRNA 3Ј-UTR used in this experiment contains four copies of the AU-rich ciselement (11), this result demonstrates the sequence-specific binding of protein(s) to the cis-element. RNA gel shift assays were performed with 30 g of cytosolic protein extract from BALB/c 3T3 cells, 10T 1 ⁄2 mouse fibroblasts, C1 and C2 mouse fibrosarcoma cells derived from 10T 1 ⁄2 cells following T24-H-ras transfection (29), human HeLa S3 cells, or COS7 monkey kidney fibroblasts. Fig. 2C shows that regardless of the origin and species of cells, the cis-element formed similar patterns of RNA-protein complex band shifts. This suggests that we have identified a widely distributed cellular protein(s) capable of interacting with the R2 3Ј-UTR cis-element.
Redox-sensitive Binding Activity of the cis-Element-There are precedents in the literature for redox-sensitive binding of protein to nucleic acid (23). We investigated the possibility that there were similar interactions with the R2 9-nt cis-element. We carried out a study of the effects of various reducing and oxidizing agents on the formation of R2BP in standard gel shift assays. Incubation of mouse BALB/c 3T3 cell extract with increasing concentrations of the reductant dithiothreitol significantly altered the cis-element binding activity (Fig. 3A, lanes  2-4). When 2-mercaptoethanol was added in increasing amounts, the intensity of the bands further increased (Fig. 3A,  lanes 5 and 6). These observations suggest that BALB/c 3T3 cell extract had some ribonucleotide reductase R2 RNA-binding protein in the oxidized form that could be reduced to achieve greater RNA binding. However, this binding activity was abolished with the preincubation of the cell extract with N-ethylmaleimide (which alkylates and thus blocks free sulfhydryl groups). Treatment of the cell extract with increasing concentrations of diamide (which catalyzes the oxidation of free sulfhydryls) inhibits formation of the cis-element-protein complex. This inhibition by diamide was dose-dependent (Fig. 3B, lanes  2-4). However, inhibition produced by 100 mM diamide was slightly reversed by 2% 2-mercaptoethanol, whereas the bind- ing activity was completely restored by 2% 2-mercaptoethanol, after preincubation of cytosolic extract with 50 mM diamide (Fig. 3B, lanes 5 and 6).
Effect of Mutations in the cis-Element Binding Site-We noted that computer-generated analysis of a 77-nt base sequence (nt positions 1225-1302) of the R2 3Ј-UTR mRNA containing the 9-nt cis-element predicts one possible structure (Fig. 4B). The binding site is along one shaft of a stem-loop in the secondary structure. The effects of various mutations in the 9-nt 5Ј-UCGUGUGCU-3Ј cis-element protein binding site were assessed by monitoring the formation of the R2BP in standard gel shift assays. Mutations within the cis-element of mouse R2 3Ј-UTR mRNA (bases 1245-1253) (30) demonstrated that certain disruptions completely prevented R2BP formation (Fig.  4C). For example, mutation of any one of the core U residues to an A residue completely abolished binding activity, but mutation of the flanking U residues to A, G, or C residues did not prevent binding activity. In addition, mutations of C residues to G residues and vice versa within this motif abolished binding activity (Fig. 4C). These results are consistent with UV crosslinking data. For example, results from UV cross-linking analyses showed that the 9-nt cis-element (5Ј-UCGUGUGCU-3Ј) formed a 45-kDa R2BP (Fig. 4D) but not the mutant fragment, 5Ј-UCGAGAGCU-3Ј (Fig. 4D). Furthermore, to examine cytosolic protein loading, a 3Ј-end-labeled human GM-CSF 3Ј-UTR (20-nt) fragment (12) was added to the binding reaction mixtures. The results from the cross-linking analyses showed that RNA-protein complexes with migration bands of 30 and 35 kDa were cross-linked to the GM-CSF 3Ј-UTR, as previously reported (12) and that the loading in the various lanes appeared to be approximately equal, as determined by densitometry (Fig. 4D).
Binding of the R2BP Correlates with Reduced mRNA Stability-To assess directly the functional role of the R2BP in the degradation of R2 mRNA, a 154-base pair R2 DNA PCR fragment (nt positions 1224 -1378) carrying the wild-type cis-element (R2/N) or a mutant fragment (R2/M) containing a series of G 3 A mutations within the binding site that eliminates binding (Fig. 5, A and B), were inserted next to the CAT coding gene (Fig. 5C). Plasmid DNAs carrying these hybrid genes (CAT/R2N, CAT/R2M), and pCAT were transfected into mouse BALB/c 3T3 cells. These transfectants were exponentially grown, followed by the addition of actinomycin D (10 g/ml) to inhibit transcription. The half-lives of the corresponding transcripts were determined as we have previously described (8 -10). The results showed that the CAT hybrid mRNA containing the normal cis-element binding site exhibited a half-life markedly shorter than that of the mRNA with the mutated insert. For example, the half-life of CAT/R2M was increased about 3-fold relative to CAT/R2N (Fig. 5C). The half-life of CAT/R2N mRNA was 1.6 h compared with that of CAT/R2M mRNA of FIG. 4. Effect of mutation in the cis-element on R2BP formation. A, the R2 mRNA binding site for the 45-kDa R2BP. B, computerpredicted structure of a 77-base segment of the R2 3Ј-UTR containing the cis-element indicated by boldface nucleotides. Bonds between complementary bases are indicated by straight lines, whereas relatively weaker bands are represented by dots. C, results from gel shift analyses of normal and mutant RNA probes. Gel shift analyses were performed using cytoplasmic extracts from mouse BALB/c 3T3 cells and the radiolabeled 9-nt cis-element or its mutant derivatives. The nucleotides in boldface type are those changed from the wild-type sequence. D, RNA UV cross-linking analyses using 3Ј-end-labeled probes (10 ϫ 10 4 cpm). Lane 1, human GM-CSF 3Ј-UTR (20 nt) as a control for cytosolic protein loading. Lane 2, 3Ј-UTR (20 nt) of the GM-CSF plus wild-type 9-nt sequence. Lane 3, 3Ј-UTR (20 nt) of the GM-CSF plus mutant 9-nt R2 cis-element. E, the consensus cis-element sequence for optimal R2BP formation, where N denotes an A, G, U, or C residue. 4.9 h. As expected, there was no significant difference between the half-life of the CAT transcript (5.3 h) and CAT/R2M (4.9 h) (Fig. 5C). These results demonstrated an inverse relationship between R2BP formation and in vivo message stability.
A Phosphorylation Signal Pathway Is Involved in the Regulation of the cis-Element Binding Activity-TPA treatment decreases the binding activity of the R2BP (8). Since TPA is a potent stimulator of protein kinase C and a tumor promoter (31,32), we tested the hypothesis that a phosphorylation pathway is critically involved in mediating the TPA effects on the R2 3Ј-UTR mRNA binding activity. Cells were treated for various times with TPA or okadaic acid, whereas the control experiments included the treatment of cells with Me 2 SO. Unlike phorbol ester tumor promoters, which activate protein kinase C, okadaic acid specifically inhibits the phosphoserine/phosphothreonine protein phosphatases 1 and 2A (32). The gel shift experiments showed that TPA treatment dramatically decreased the cis-element binding activity with time (Fig. 6A), in agreement with previous studies (8). Interestingly, similar to the TPA effect, okadaic acid treatment also significantly decreased the binding activity (Fig. 6B). Moreover, in the control experiments, Me 2 SO had no effect on the RNA-binding activity (Fig. 6C). These results suggest that a relationship exists between the increase in cellular protein phosphorylation in vivo and the R2 cis-element binding activity. If TPA and okadaic acid altered RNA-protein binding activity and message stability properties, presumably through a phosphorylation signal pathway, then their effects should be concentration-dependent. Fig. 6, D and E, shows that this is the case. For example, increasing the concentration of TPA and okadaic acid led to corresponding reductions in the 9-nt cis-element binding activity. On the other hand, increasing the concentration of Me 2 SO to as high as 50 M did not affect protein binding to the cis-element (Fig. 6F). DISCUSSION We have previously shown that the 3Ј-UTR is implicated in the destabilization of R2 mRNA and contains TPA-responsive elements that may be associated with R2 message stabilization (8). We have identified a putative novel 9-nt cis-element, 5-UCGUGUGCU-3Ј in the 3Ј-UTR of R2 mRNA, which interacts with a cytosolic protease-sensitive factor(s) to form a 45-kDa R2BP. The binding activity of the cis-element was sequence-specific. Similar to several previously published mRNA cis-elements (33), there is only one copy of the 9-nt cis-element within the R2 message. The cis-element binding site showed a similar band shift complex in a variety of cells of different species, suggesting a common pathway for the regulation of R2 mRNA degradation. Mutagenesis studies indicated that there are relatively stringent primary sequence requirements for formation of the R2BP.
Our results indicate that the cis-element binding activity is redox-sensitive, because formation of the cis-element-protein complex was abolished by treatment of cell extracts with Nethylmaleimide and diamide; however, the inhibition of RNAprotein complex formation by diamide was reversed by 2% mercaptoethanol, suggesting a possible sulfhydryl switch mechanism. The present study connects, for the first time, the action of the tumor promoters TPA and okadaic acid and phosphorylation signal-controlled pathway(s) in a mechanism(s) leading to message stability that is associated with down-regulation of the R2BP activity. As expected, treatment of cells with okadaic acid, which exerts its effects by causing a net increase in the prevailing levels of phosphorylated proteins, resulted in the down-regulation of the cis-element binding activity. Such a mechanism has been proposed for the posttranscriptional regulation of ribonucleotide reductase R1 mRNA stabilization (20). Our data indicated that insertion of a PCRamplified DNA fragment containing the wild type protein binding site into a heterologous CAT gene caused message destabilization, whereas mutations introduced into the normal ciselement that eliminated formation of the R2BP had no significant effect on the stability of the CAT hybrid RNA. This result is consistent with the idea that the in vivo message destabilization of the R2 mRNA is mediated by the R2BP in cells, and down-regulation of the binding activity of this complex leads to a decrease in the degradation of R2 mRNA. We conclude that the cis-element functions as a destabilizing sequence.