Báo cáo khoa học: Sequences downstream of the transcription initiation site are important for proper initiation and regulation of mouse ribonucleotide reductase R2 gene transcription ppt

Eur J Biochem 270, 1791–1801 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03541.x Sequences downstream of the transcription initiation site are important for proper initiation and regulation of mouse ribonucleotide reductase R2 gene transcription Irina Kotova, Anna L Chabes*, Sergei Lobov, Lars Thelander and Stefan Bjorklund ă Department of Medical Biochemistry and Biophysics, Umea˚ University, Sweden Ribonucleotide reductase is essential for the synthesis of all four dNTPs required for DNA replication The enzyme is composed of two proteins, R1 and R2, which are both needed for activity Expression of the R1 and R2 mRNAs is restricted to the S-phase of the cell cycle, but the R1 and R2 promoters show no obvious sequence homologies that could indicate coordination of transcription Here we study initiation of transcription at the natural mouse R2 promoter, which contains an atypical TATA-box with the sequence TTTAAA, using a combination of in vivo reporter gene assays and in vitro transcription Our results indicate that in constructs where sequences from the R2 5¢-UTR are present, the mouse R2 TATA-box is dispensable both for unregulated, basal transcription from the R2 promoter and for S-phase specific activity Instead, initiation of R2 transcription is directed by sequences downstream from the transcription start We report that this region contains a conserved palindrome sequence that interacts with TAFIIs This interaction down-regulates basal transcription from the R2 promoter, both in the absence and in the presence of the TATA-box Efficient transcription initiation at a eukaryotic proteinencoding gene requires assembly on promoter DNA of a protein complex containing RNA polymerase II and five general transcription factors (GTFs) IIB, IID, IIE, IIF, and IIH (reviewed in [1–3]) TFIID refers to a multiprotein complex composed of the TATA-binding protein (TBP) and a set of proteins called TBP-associated factors (TAFs) Transcription of most eukaryotic genes is initiated around 20–30 base pairs downstream from a conserved sequence called the TATA-box, which binds TBP as a first step in a sequence of events leading to formation of a functional preinitiation complex [4] The TATA-box sequence is conserved through evolution [consensus sequence: TATA(A/T)A(A/T)] but it is known that TBP can interact with sequences that differ considerably from this consensus sequence However, at most natural promoters, TBP is not solely responsible for promoter recognition Rather, a more extended sequence around the TATA-box interacts with both TAFs and other GTFs, such as TFIIB Mapping of interactions between the human TFIID subunits and the adenovirus major late promoter using a photocrosslinking method showed that TFIID interacts with promoter sequences both upstream and downstream of the TATAbox, but also that TFIID–DNA interactions are formed downstream from the position for transcription initiation [5] In contrast to the TATA-box, these upstream and downstream sequences show no obvious homology when different promoters are compared Regulation of transcription is normally explained by signaling from regulatory proteins binding to specific promoters DNA sequences located upstream from the TATA-box These signals are then transferred to the general transcription machinery via coactivators or corepressors such as the Mediator complex [6,7] However, considering that the sequences surrounding the TATAbox, which interact with components of the general transcription machinery, show little or no homology between different promoters, it is likely that transcription is also regulated at this level This type of regulation would then both be dependent on the strength of interactions between different promoter DNA and the GTFs depending on the sequence surrounding the TATA-box, but also by formation of preinitiation complexes that contain different general transcription factors For example, it is known that TAF-complexes exist in multiple forms, some even lacking TBP [8,9] It is therefore not surprising that expression from specific promoters is regulated by recruitment of different TAF-complexes for example in different tissues, at different developmental stages or at different phases of the cell cycle [reviewed in 10] Furthermore, the fact that TBP is also required for transcription of genes that lack a TATA-box, shows that the requirement of TBP for initiation is uncoupled from the requirement of a consensus TATA-box Correspondence to S Bjorklund, Department of Medical Biochemistry ă and Biophysics, Umea University, SE-901 87 Umea, Sweden Fax: +46 907869795, Tel.: + 46 907866788, E-mail: stefan.bjorklund@medchem.umu.se Abbreviations: 5¢-UTR, 5¢-untranslated region; GTFs, general transcription factors; TBP, TATA-binding protein; TAFs, TBP associated factors; DPE, downstream promoter element; Py, pyrimidine; Im, imidazole *Present address: Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11724, USA (Received 23 December 2002, revised 21 February 2003, accepted 26 February 2003) Keywords: in vitro transcription; ribonucleotide reductase; TAFs; TATA-box; transcription regulation Ó FEBS 2003 1792 I Kotova et al (Eur J Biochem 270) Ribonucleotide reductase catalyzes the formation of deoxyribonucleotides from the corresponding ribonucleotides, which is the rate-limiting step in the production of precursors for DNA synthesis [11] The mouse ribonucleotide reductase is composed of two subunits, proteins R1 and R2 Both subunits are required for enzyme activity and it could be reasonable to assume that the expression of the R1 and R2 mRNAs should be coordinated at the transcription level Accordingly, both the R1 and R2 mRNAs show an S-phase specific expression with low or undetectable levels in G1 cells, and a synchronized increase as cells enter the S-phase of the cell cycle [12] However, analysis of the mouse R1 and R2 promoter sequences show that they differ, not only from each other, but also from the common promoter structure found in the majority of eukaryotic genes The mouse R1 promoter lacks TATA-box and instead contains an initiator sequence that has been shown to interact with the transcription factor TFII-I [13], and a downstream sequence called c that shows no homology to previously identified downstream promoter elements [14] The protein(s) that interacts with the c sequence is so far unidentified In contrast, the mouse R2 promoter contains no initiator element-sequence or c element, but instead has an atypical TATA-box with the sequence TTTAAA, located approximately 30 base pairs upstream from the mapped major transcription start sites in the R2 promoter [15] The transcriptional efficiency for this variant of the TATA-box has been studied in the adenovirus major late promoter context [16] Compared to the consensus TATA-box sequence, the R2-type sequence resulted in a 75% reduction in transcriptional efficiency However, no analysis of how these mutations affected the transcription start position was presented Here we examine the mouse ribonucleotide reductase R2 TATA-box and sequences downstream of it for their function in transcription from the R2 promoter Our data indicate that the R2 TATA-box is dispensable for basal and S-phase-specific expression from the R2 promoter, but that it is important for the position of transcription initiation However, the R2 TATA-box is required for maximal promoter strength in the presence of sequences downstream from the R2 transcription start and the availability of proteins binding to these sequences Experimental procedures Plasmids The internal control plasmid pML311 contains a shorter (311 bp) G-less cassette fused with the adenovirus major late promoter It was created from pML(CAT)19 [17] by PCR using one primer complementary to the polylinker upstream from the AdML promoter and a second primer complementary to nucleotide 311–292 bp downstream from the start of the G-less sequence, followed by cleavage of the PCR product with EcoRI and BamHI and ligation into pML(CAT)19 digested with the same enzymes The R2-luciferase reporter construct (pAC 10, here also called TATAWt UTRWt) and the R2 promoter construct with mutation in the CCAAT box have been described previously [18] The R2 TATAmut UTRWt-luciferase reporter construct is analogous to the pAC10/TATAWt UTRWt except that the TATA-box (TTTAAA) at the position )29 bp relative to the mouse R2 transcription start site was substituted by the sequence GCGCGC by overlap extension PCR The TATAWt UTRmut and the TATAmut UTRmut constructs fused to the G-cassette were made by overlap extension PCR using pML(CAT)19 and the pAC10/ TATAWt UTRWt or the TATAmut UTRWt-luciferase reporter constructs as templates Creation of a TATAWt UTRWtG-less cassette construct with exactly the same 5¢-UTR sequence as the analogous luciferase construct failed probably because primers specific for the mouse R2 5¢-UTR sequence must include a 10-bp perfect palindrome which might cause secondary primer structures Instead we used primers further downstream from the R2 transcription start site, nucleotides +43 to +22 relative to the major transcription start site and introduced an NcoI site between the end of the R2 5¢-UTR and the G-less cassette The TATAmut UTRWt-G-less cassette construct was created from the TATAmut UTRWt-luciferase construct in the same way as the TATAWt UTRWt-G-less cassette construct Therefore, these constructs contain a longer sequence from the R2 5¢UTR compared to the corresponding luciferase constructs (42 instead of 17 nucleotides) The TATAWt UTRmut- and the TATAmut UTRmut-luciferase reporter constructs were made from the corresponding G-less cassette constructs As these constructs, in contrast to the corresponding G-less constructs need to be efficiently translated, we decided to replace the R2 5¢-UTR with a nonrelated 5¢-UTR instead of deleting it Therefore, these constructs were made by amplification of the R2 promoter from corresponding G-less cassettes constructs using a downstream primer that included 23 nucleotides from the luciferase 5¢-UTR Synthesis of the mouse R2-specific polyamide The R2 TATA-specific Py-Im polyamide (Fig 1A) was synthesized by solid-phase methods as described [19] The purity and identity of the polyamide was verified by analytical high-pressure liquid chromatography, 1H nuclear magnetic resonance, and matrix-assisted laser desorption ionization-time of flight mass spectrometry The polyamide was dissolved in distilled water and the concentration of the stock solution was calculated using an extinction coefficient of e ¼ 78 000 M)1Ỉcm)1 at 310 nm The binding affinity of the polyamide to the R2 promoter was determined to Kd ¼ nm by quantitative DNase footprinting under equilibrium conditions [20] Treatment of stably transformed cells with polyamide Balb/3T3 cells (5 · 105 per 10 cm dish) stably transformed with the pAC10 full-length R2 promoter-luciferase reporter gene construct [21] were grown overnight in 10 mL Dulbecco’s modified Eagle’s medium (DMEM) + 10% heat-inactivated horse serum After 24 h, and 50 lM R2 polyamide (in DMEM) was added to the plates After h, cells were harvested and assayed for luciferase activity as described [21] Gel shift experiments Nuclear extracts used in gel shift experiments with TBP were prepared from logarithmically growing Balb/3T3 cells Ĩ FEBS 2003 Role of RNR2 5¢-UTR in transcription initiation (Eur J Biochem 270) 1793 [22] Gel shift experiments were made using the oligonucleotides 5¢-GCGGTTGGGTGGCTCTTTAAAGGGCG CG-3¢ and 5¢-CGCGCCCTTTAAAGAGCCACCCAA CCGC-3¢ In oligonucleotides with mutated TATA-box the sequence TTTAAA (shown in bold) is substituted by GCGCGC Single-stranded oligonucleotides were endlabeled using T4 polynucleotide kinase and [c-32P]ATP (specific activity 3000 CiỈpmol)1, Amersham Biosciences) as previously described [23], annealed by heating to 65 °C followed by cooling down slowly to the room temperature and purified by gel filtration on Sephadex G-50 (Amersham Biosciences) A typical binding reaction contained fmol labeled oligonucleotide and 10 ng (26 fmol) recombinant human TBP (Promega) in 10 lL of binding buffer [10% glycerol, 20 mM Tris/HCl (pH 8.0), 80 mM KCl, 10 mM MgCl, mM dithiothreitol] After 15 incubation at room temperature, lL loading buffer [40% glycerol, 250 mM Tris/HCl (pH 8.0)] was added to the reactions and DNAprotein complexes were resolved by electrophoresis through 6% polyacrylamide gels in electrophoresis buffer (0.5 · Tris/borate/EDTA, mM MgCl2, 0.02% NP-40) The gels were run at +4 °C, dried and subjected to autoradiography Nuclear extracts used in gel shift experiments with the palindrome sequence in the R2 5¢-UTR were prepared from Ehrlich–Lettre ascites mouse carcinoma cells (ATCC No CCL77) as previously described [24] The gel shift experiments were performed essentially as described in [25] using the oligonucleotides: wild-type: 5¢-CAGTCGGCGGTGC ACCGGATTCCAGCTGTTT-3¢; mutation 1: 5¢-CAGT CGGCTGTCAACCGGATTCCAGCTGTTT-3¢; muta- Fig A polyamide specific for the mouse R2 TATA-box interferes with TBP-binding to the R2 promoter in vitro but has no effect in vivo (A) Structure of the Im-b-ImPyPy-c-ImImPyPyPy-b-Dp R2 TATA-specific polyamide and the sequence in the mouse R2 promoter to which it binds Im-Py targets GỈC base pairs and Py-Im targets CỈG base pairs and the Py-b-alanine pair recognizes both T and TỈA base pairs [38] The boxed sequence represents the atypical R2 TATA-box Black and white circles, Im and Py rings, respectively; curved line, hairpin junction, which is formed with c-aminobutyric acid; diamonds, b-alanine; parenthesis with plus sign, 3-(dimethylamino)propylamine (B) Gel shift experiments with a 32P-labeled oligonucleotide representing the R2 TATA-box Lane 1, only labeled oligonucleotide; lane 2, labeled oligonucleotide incubated with 10 ng (26 fmol) recombinant human TBP; lanes 3–6, as in lane but with decreasing amounts (0.15 pmol, 0.075 pmol, 0.03 pmol, 0.0075 pmol respectively) of unlabeled R2 TATA-box oligonucleotide added to the binding reactions Lanes and 8, as lanes and 2, respectively; lanes 9–11, as lane but with increasing amounts of polyamide (540 fmol, 3.38 pmol, 33.8 pmol, which is a 20-, 130-, and 1300-fold excess compared to TBP, respectively) added to the binding reactions Arrows to the right indicate the position for free probe (lower) and the TBP-probe complex (upper) Numbers below the figure represent quantifications of the TBP-probe complexes in each lane The quantifications were made using the Scion Image software (C) Addition of the R2-polyamide to Balb/3T3 cells, which are stably transfected with a reporter gene construct where the wild-type R2 promoter controls the luciferase gene (TATAWtUTRWtluciferase) tion 2: 5¢-CAGTCGGCGGTGCACCGTAATCCAGCT GTTT -3¢; mutation 3: 5¢-CAGTCGGCGGTGCACCG GATTCCGAGTGTTT-3¢ Nucleotides in bold represent mutated nucleotides and the palindrome sequence is underlined After labeling and annealing, the probes were Ó FEBS 2003 1794 I Kotova et al (Eur J Biochem 270) purified by electrophoresis in 5% polyacrylamide gel under nondenaturing conditions In competition assays, the annealed unlabeled oligonucleotides were mixed with labeled probe prior to the addition of nuclear extracts In experiments with antibodies, antibodies were preincubated with nuclear extracts for 15 on ice in the binding buffer before adding the labeled probe 65 °C followed by slow cooling to room temperature After the primer extension reaction, the products where denatured for at 95 °C the reaction products were resolved by electrophoresis on M urea, 10% polyacrylamide gels and subjected to autoradiography Transfection of cells, serum starvation and luciferase assays TBP binding to the mouse R2 promoter TATA-box is not required for transcription from the R2 promoter in vivo Transient transfection, isolation of stably transformed cells, synchronization of Balb/3T3 cells by serum starvation and luciferase assays were done as described previously [18] The relative luciferase value was determined as firefly luciferase activity normalized against Renilla luciferase activity multiplied by 1000 Preparation of nuclear extracts and in vitro transcription assays Nuclear extracts were prepared from logarithmically growing Ehrlich–Lettre ascites cells as previously described [24] In vitro transcription assays with purified transcription factors were performed as described in [18] In vitro transcription assays with crude nuclear extracts were made in a similar way but with the following modifications Each transcription reaction contained a total amount 35–45 lg of nuclear extract For experiments using templates containing the mutated 5¢-UTR upstream from the G-less cassette, the · transcription mixture contained 1.6 mM GTP in addition to the other three nucleotides Stop mixture contained 10 mM Tris/HCl (pH 7.5), 0.3 M NaCl, mM EDTA, 0.1 mgỈmL)1 glycogen (Boerhinger Mannheim) and 130 mL)1 T1 ribonuclease After incubation for 30 at 25 °C the reaction was stopped by addition of 200 lL stop mixture and incubated for 30 at 37 °C prior to proteinase K treatment When an internal control was included in the reactions, each reaction contained 60 fmol of the DNA template to be studied, and 60 fmol of pML 311 In the experiments with TAFII135 antibodies, the antibodies were added to the reaction mixtures at different concentrations and the reactions were incubated on ice for 15 prior to addition of the 2· transcription mixture Primer extension Primer extension reactions were carried out essentially as described in [26] Avian myeloblastosis virus reverse transcriptase, T4 polynucleotide kinase and /X174 DNA/HinfI dephosphorylated marker were purchased from Promega RNAs to be used as template for the primer extension reactions were synthesized by in vitro transcription reactions using nuclear extracts and the different R2 promoterluciferase constructs The oligonucleotide used for primer extension was complementary to the nucleotides 65–90 in the coding strand of the luciferase cDNA (5¢-CTCTTCATA GCCTTATGCAGTTGCTCTCCAG-3¢) The labeled oligonucleotide (0.1 pmol, specific activity  · 106 cpmỈpmol)1) was added to the in vitro transcribed RNA, the volume was adjusted to 20 lL with water and the oligonucleotide was annealed to the RNA by heating to Results In order to study initiation of transcription from the R2 promoter, we obtained a polyamide specific for a sequence at the 5¢-end of, and immediately upstream from, the atypical mouse R2 TATA-box (TTTAAA; Fig 1A) DNase1 footprinting analysis showed that the synthesized polyamide bound specifically to this sequence with a Kd of nM Gel shift experiments using an end-labeled oligonucleotide corresponding to the region around the R2 TATA-box (Experimental procedures), recombinant human TBP and the polyamide showed that human TBP binds specifically to the mouse R2 TATA-box and that the polyamide interferes with this binding (Fig 1B) To study if the polyamide also inhibits transcription from the R2 promoter in vivo, polyamide was added to Balb/3T3 cells stably transfected with a reporter gene construct where the luciferase gene is under the control of the full-length mouse R2 promoter [21] To our surprise we found that the polyamide had no effect on R2-promoter-luciferase expression even when cells are incubated with high concentrations (50 lM) of polyamide (Fig 1C) Importance of the R2 TATA-box and TBP binding to the R2 promoter We next studied the importance of TBP and the R2 TATA-box in an in vitro transcription system reconstituted from recombinant mouse TBP, TFIIB, TFIIE, TFIIF and highly purified mouse TFIIH and RNA polymerase II [18] As templates for these experiments we used constructs where a G-less cassette is ligated directly downstream from the nucleotide that has been mapped as the major transcription initiation site in the mouse R2 gene The only difference between the two templates is that they either contain the natural R2 TATA-box or a mutation of it to GCGCGC (Fig 3D; G-less templates TATAWtUTRmut and TATAmutUTRmut) We found that both TBP (Fig 2A, compare lanes and 5) and the R2 TATA-box (Fig 2A, compare lanes and 3) were absolutely required for transcription from the R2 promoter in this system As a positive control for these experiments, we included a reaction using a template where the G-less cassette is controlled by the adenovirus major late (AdML) promoter (lane 1) We found that the activity of the natural R2 promoter is 64% of the AdML promoter strength in this basal, unregulated in vitro transcription system (Fig 2A, compare lanes and 2) In contrast to the results obtained in vitro, analysis of the importance of the R2 TATA-box in vivo using the R2-luciferase reporter genes showed that a mutation of the R2 TATA-box only resulted in a limited decrease of the Ĩ FEBS 2003 Role of RNR2 5¢-UTR in transcription initiation (Eur J Biochem 270) 1795 R2 promoter strength (Fig 2B) Finally, we also performed transient transfection experiments in synchronized Balb/3T3 cells using full-length R2 promoter-luciferase constructs with wild-type or mutated TATA-box (Fig 2C; TATAWtUTRWt and TATAmutUTRWt) These results showed that a mutation of the R2 TATA-box does not effect the S-phase specific expression from the mouse R2 promoter The discrepancy between the in vitro results described in Fig 2A, which show that the R2 promoter is dependent both of TBP and its TATA-box, and the in vivo results presented in Figs 1C and 2B, which indicate that the R2 TATA-box has a very limited effect on transcription from the R2 promoter, could possibly be explained by the lack of an essential transcription factor in the reconstituted in vitro transcription system In order to study transcription from the mouse R2 promoter in vitro in a more complex context, we therefore used an in vitro transcription system based on a crude nuclear extract In line with the results obtained using the defined in vitro transcription system reconstituted from pure general transcription factors, also this crude system showed that the mouse R2 promoter is highly dependent on its TATA-box for full transcription (Fig 3A, compare lanes and 2) As indicated, transcription from the TATAmutated R2-promoter construct is fourfold lower compared to the natural R2 promoter Sequences downstream of the R2 transcription initiation site are important for proper initiation of mouse R2 transcription Fig Both the R2 TATA-box and TBP are required for transcription from the R2 promoter in a defined basal in vitro transcription system but the R2 TATA-box is dispensable for basal and S-phase specific transcription from the R2 promoter in vivo (A) In vitro transcription experiments using highly purified mouse GTFs and the G-less cassette reporter gene under the control of: lane 1, the AdML promoter; lane 2, the full-length mouse R2 promoter (TATAWtUTRmut); lane 3, the full-length R2 promoter with a mutation in the TATA-box (TATAmutUTRmut); lane 4, same as lane 2; lane 5, same as lanes and but with TBP omitted from the in vitro transcription reaction (B) Transient transfection experiments using the indicated amounts of the full-length mouse R2 promoter (TATAWtUTRWt) ligated to the luciferase gene (black bars) and of the full-length R2 promoter with a mutation in the TATA-box (TATAmutUTRWt), ligated to the luciferase gene (white bars) (C) A mutation of the TATA-box does not effect the S-phase specific expression from R2 promoter Balb/3T3 cells were transiently transfected with the full-length R2 promoter-luciferase constructs, TATAWtUTRWt (m) and TATAmutUTRWt (j) The graph shows relative luciferase values at the indicated time points after release from serum starvation Our results so far indicate a difference in the requirement of the R2 TATA-box when comparing results obtained in vitro and results obtained in vivo This could still reflect a difference between the two systems, either because the in vitro systems lack an essential transcription factor or because of interactions between the template and chromatin components in the in vivo system However, also the templates used for the in vitro transcription and in vivo luciferase experiments differed from each other Our wildtype R2-luciferase template (TATAWtUTRWt-luciferase) contains a sequence from the mouse R2 5¢-UTR while the corresponding wild-type R2 promoter G-less template (TATAWtUTRmut-G-less) lacks this sequence In order to study if this difference was important, we made additional luciferase reporter gene constructs where the native or TATA-mutated R2 promoters were fused to an unrelated 5¢-UTR (Fig 3D summarizes all template constructs) We then tested the new luciferase constructs in vivo in transient transfection experiments Comparison of these two new R2-luciferase constructs, which contain a mutated 5¢-UTR, shows that they differ in expression levels in a way that indicates a requirement for the R2 TATA-box also in vivo (Fig 3B) As shown, a mutation of the TATA-box in combination with a mutated 5¢-UTR causes a threefold reduction in transcription also in vivo In conclusion, the experiments presented show that the R2 TATA-box is important for initiation both in vitro (Fig 3A) and in vivo (Fig 3B) when the R2 5¢-UTR is either deleted or mutated In contrast, the in vivo results suggest that when the R2 5¢-UTR is included in the template, the R2 TATA-box becomes redundant 1796 I Kotova et al (Eur J Biochem 270) Ó FEBS 2003 Fig Importance of the R2 TATA-box for promoter activity in the absence of the R2 5¢-UTR (A) In vitro transcription experiments using the R2-promoter G-less templates TATAWtUTRmut (lane 1) and TATAmutUTRmut (lane 2) (described in Fig 3D) Two templates were used in each experiment: the R2 promoter construct ligated to a longer G-less cassette (product indicated by the upper arrow to the left of the autoradiograph) and the adenovirus major late promoter ligated to a shorter G-less cassette (product indicated by the lower arrow to the left of the autoradiograph) All bands were quantified using the Scion Image program and the ratio between the upper and lower bands for each assay is presented below the autoradiograph (B) Transient transfection experiments with increasing amounts of DNA Black bars: the TATAWtUTRmut R2-luciferase construct; white bars: the TATAmutUTRmut R2-luciferase construct The values on the y-axis indicate the relative luciferase values after normalization of the R2 promoter luciferase to the cotransfected SV40-driven Renilla luciferase values for each transfection experiment (C) In vitro transcription experiments using the R2-promoter G-less templates 1, TATAWtUTRWt; 2, TATAmutUTRWt (described in Fig 3D) The numbers below the autoradiograph represent the ratio between the upper and lower bands (D) Overview of the reporter gene constructs used in experiments All constructs start at the Pvu II restriction site located at nucleotide )1500 relative to the R2 transcription start site TATAWt indicates that the construct contains the wild-type R2 TATA-box, TATAmut indicates that the construct carries a TTTAAA fi GCGCGC mutation, UTRWt indicates that the construct contains 17 base pairs (luciferase constructs) or 42 base pairs (G-less cassette constructs) from the wild-type mouse R2 5¢-UTR Finally, UTRmut indicates that the construct either lacks a 5¢-UTR (the G-less cassette constructs) or that it contains a 5¢-UTR composed of 21 nucleotides from the 5¢-end of the G-less cassette fused to 18 nucleotides from the 5¢-UTR of the luciferase reporter gene Arrows indicate the corresponding position for the wild-type R2 transcription start site (E) A mutation of the R2 5¢-UTR does not effect the S-phase specific expression from R2 promoter Balb/3T3 cells were transiently transfected with the full-length R2 promoter-luciferase constructs, TATAWtUTRWt (d) and TATAWtUTRmut (r) The graph shows relative luciferase values at the indicated time points after release from serum starvation results from these experiments confirm the in vivo results As seen in Fig 3C, we found that transcription from these templates, which include the R2 5¢-UTR, was almost independent of the R2 TATA-box (compare to Fig 2B) Similarly to the experiments with the mutation of the R2 TATA-box (Fig 2C), we also performed transient transfection experiments in synchronized Balb/3T3 cells using full-length R2 promoter-luciferase constructs with wildtype or mutated 5¢-UTR (Fig 3E TATAWtUTRWt and TATAWtUTRmut) We found that a mutation of the R2 5¢-UTR had no effect on the S-phase specific expression from the mouse R2 promoter In reciprocal experiments we also made new G-less templates for in vitro transcription that included the R2 5¢-UTR and either contained the natural R2 TATA-box or a mutation of it to GCGCGC (G-less templates TATAWtUTRWt and TATAmutUTRWt in Fig 3D) The Mutation of the R2 TATA-box affects the position for transcription initiation It was possible that the mutation of the TATA-box and the different 5¢-UTR could result in a change of transcription start that might in turn influence the observed transcription levels We therefore performed primer extension assays on Ó FEBS 2003 Role of RNR2 5¢-UTR in transcription initiation (Eur J Biochem 270) 1797 Protein interaction with the R2 5¢-UTR Fig Determination of the transcription start site in the different R2-luciferase constructs The indicated R2 promoter luciferase constructs were used as templates in in vitro transcription experiments and the resulting RNA products were used for primer extension experiments The same primer, specific for the luciferase open-reading frame was used for all constructs Arrows indicate the position for the two major transcription start sites for each template RNA synthesized in vitro using all four luciferase constructs in nuclear extract-based transcription assays The luciferase templates were used in these experiments as it was difficult to find suitable sequences for synthesis of a primer in the G-less cassette Similarly to the previously reported mapping of the transcription start in the wild-type R2 gene [15], we found that all four constructs used here initiate transcription on two adjacent nucleotides (Fig 4) We also found that both constructs that contain the wild-type TATA-box (TATAWtUTRWt and TATAWtUTRmut) initiate transcription at the positions used by the native R2 gene In contrast, the two constructs that contain the mutated TATA-box (TATAmutUTRWt and TATAmutUTRmut) both initiate transcription 2–3 base pairs upstream from the normal R2 promoter initiation sites However, this change in transcription start site does not correlate to the expression levels from the different promoters (compare to Figs 3A–C and 4) Please observe that the UTR of the UTRmut templates is longer than the UTRWt templates, which explains the differences in length of the primer extension products between these two types of constructs Our data presented above suggest that the R2 5¢-UTR contains sequences that bind protein(s) which assist TBP in the formation of a functional preinitiation complex However, the R2 5¢-UTR showed no homologies to other 5¢-UTR sequences that previously have been identified as important for expression from different promoters, for example downstream promoter element (DPE) [14] We therefore analyzed the sequence of the R2 promoter and the 5¢-UTR, and compared it to the corresponding region in the human R2 gene We could identify a potentially interesting palindrome in the R2 5¢-UTR sequence that covers 10 base pairs and overlaps with the position for transcription initiation in the mouse R2 promoter (Fig 5A, palindromes in boxes) Interestingly, this sequence is also conserved in the human promoter except for the second and the last base pairs In general, the sequence from the conserved atypical TATA-box into the first 30 nucleotides of the 5¢-UTR shows a much higher homology between mouse and human (66% identity) compared to either the sequence upstream or downstream from this sequence To study if the sequence downstream from the mouse R2 transcription start site interacts with proteins, we performed gel shift assays using a labeled oligonucleotide corresponding to base pairs )8 to +23 in the R2 gene relative to the major transcription start site, and nuclear extracts prepared from logarithmically growing Ehrlich–Lettre ascites mouse cells Incubation of nuclear extract with this oligonucleotide resulted in formation of a DNA-protein complex (Fig 5B, lanes and 4), which could be competed specifically by an unlabeled oligonucleotide with the same sequence as the labeled oligonucleotide (Fig 5B, lane 3) However, it could not be competed by unrelated oligonucleotides (data not shown) We also performed similar experiments using an oligonucleotide that included the R2 TATA-box (nucleotides )34 to +23) The results from these experiments were identical to the results obtained with the shorter oligonucleotide (data not shown) TAF subunits have been shown to interact to sequences both upstream and downstream from the TATA-box of promoters In addition, mapping of interactions between different TAF subunits and the adenovirus major late promoter also showed interactions between TAF250 and TAF135 and sequences even downstream from the transcription start [5] We therefore wanted to study if the sequence downstream from the R2 transcription start site could interact with TFIID, but all commercially available TAF antibodies are specific for human TAFs and show no cross-reactivity with the corresponding mouse proteins according to the manufacturers We could however, obtain a monoclonal antibody that recognizes mouse TAFII135 in Western blots and which also can immunoprecipitate the mouse TFIID-complex (W S Mohan II, E Scheer, O Wendling, D Metzger & L Tora, personal communication) Incubation of a nuclear extract with monoclonal antibodies against TAF135 abolished formation of the complex (Fig 5B, lanes 5–7), whereas incubation with the same amounts of monoclonal antibodies against the ribonucleotide reductase R1 protein had no effect (Fig 5B, lanes 8–10) 1798 I Kotova et al (Eur J Biochem 270) Ó FEBS 2003 Fig Sequences downstream from the R2 transcription start site interact with proteins and affect R2 transcription (A) Comparison of the mouse and human R2 DNA sequences around their transcription start sites Capital letters in bold style indicate nucleotides that are conserved between human and mouse; arrows indicate the mapped major transcription start at each promoter and boxes represents the partially conserved palindrome sequence described in the text (B) Gel shift experiments using nuclear extracts from logarithmically growing Ehrlich–Lettre ascites cells and a 32P-labeled wild-type oligonucleotide including the transcription start and the downstream sequence from of the mouse R2 gene The protein-DNA complex is indicated by the arrow to the right of the autoradiograph Lanes 2–10, 32P-labeled R2 5¢-UTR oligonucleotide with nuclear extracts (10 lg) from exponentially growing Ehrlich–Lettre ascites cells; lane 3, a 100-fold molar excess of unlabeled R2 5¢-UTR oligonucleotide was added; lanes 5–7, nuclear extract was preincubated with different amounts (0.43 lg, 0.85 lg and 1.7 lg, respectively) of monoclonal antibody towards human TAFII135 for 10 on ice prior to addition of the probe; lanes 8–10, the same as 5–7, but with monoclonal antibodies against mouse R1 The numbers under the gel represent the intensity of each band normalized to the band in lane (C) Gel shift experiment using nuclear extracts from logarithmically growing Ehrlich–Lettre ascites cells, wild-type and mutated oligonucleotides including the transcription start and the downstream sequence from of the mouse R2 gene Lanes 1–3: 32P-labeled R2 5¢-UTR wild-type oligonucleotide; lanes and 3, 32P-labeled R2 5¢-UTR oligonucleotide with nuclear extracts (10 lg) from exponentially growing Ehrlich–Lettre ascites cells; lane 3, nuclear extracts were preincubated with 1.7 lg of the TAFII135 monoclonal antibody; lanes 4–6, the same as lanes 1–3, but using a labeled oligonucleotide containing mutations in the palindrome sequence (mutation 1, experimental procedures); lanes 7–9, as lanes 1–3 but using a labeled mutation oligonucleotide; lanes 10–12, as lanes 1–3 but using a labeled mutation oligonucleotide (D) In vitro transcription experiments with nuclear extracts isolated from logarithmically growing Ehrlich–Lettre ascites cells, the TATAWtUTRWt R2-G-less cassette reporter gene (lanes 1–3), the TATAWtUTRmut R2-G-less cassette reporter gene (lanes 4–6) and monoclonal antiTAFII135 antibodies Lanes and 4, no antibody added to the in vitro transcription reaction; lanes and 5, lL of 300-fold diluted antibody was added to the in vitro transcription reactions; lanes and 6, lL of 60-fold diluted antibody was added to the in vitro transcription reactions All bands were quantified using the Scion Image program The upper numbers represent the intensity of each band normalized to the band in lane The lower numbers represent the intensity of the bands in lanes and normalized to the band in lane In order to map the position in the R2 5¢-UTR where the TAF135 protein interacts, we synthesized three different oligonucleotides corresponding to nucleotides )8 to +23 in the R2 gene Each oligonucleotide carried mutations in discrete regions (Experimental procedures) Gel shift experiments using these mutated oligonucleotides and nuclear extracts showed the only mutations within the palindrome sequence (mutation 1) resulted in a shift that differed from the one observed using the wild-type oligonucleotide (Fig 5C and data not shown) As seen in Fig 5C, the oligonucleotide with mutations in the palindrome sequence resulted in two gel shift bands compared to the single band observed with the wild-type oligonucleotide However, neither of these two bands could be competed by addition of TAF135 antibodies We next included the TAF135 antibody in in vitro transcription experiments using nuclear extracts and the G-less cassette constructs that contain the native R2 TATA- box and either the natural or the deleted R2 5¢-UTR (TATAWtUTRWt and TATAWtUTRmut, respectively) We found that addition of increasing amounts of the TAFII135 antibody resulted in an up to 5.1-fold increase in transcription from the TATAWtUTRWt template (Fig 5D, lanes 1–3) A corresponding, 4.8-fold increase was also found when comparing the wild-type (TATAWtUTRWt) and the template that lacks the R2 5¢-UTR (TATAWtUTRmut; compare lanes and 4) In contrast, addition of the TAFII135 antibody to reactions using the TATAWtUTRmut template resulted in a much less pronounced increase in transcription (1.6-fold, compare Fig 5D, lanes 4–6) In control experiments, monoclonal antibodies against the ribonucleotide reductase R1 protein had no effect on transcription efficiency in similar in vitro transcription Ó FEBS 2003 Role of RNR2 5¢-UTR in transcription initiation (Eur J Biochem 270) 1799 experiments (data not shown) The fact that both a deletion of the R2 5¢-UTR or addition of the monoclonal TAFII135 antibodies to the transcription reactions results in a similar fivefold increase in transcription shows that TFIID binds to the mouse R2 5¢-UTR and causes down-regulation of transcription from the R2 promoter Finally, we also performed primer extension experiments on in vitro transcription reactions were the TAFII135 antibody was included We found that inclusion of the TAFII135 antibodies in the in vitro transcription reaction had no effect on the position for the R2 transcription start site (data not shown) Discussion Previous and present experiments on the regulation of expression of the ribonucleotide reductase R2 subunit are focused on sequences located upstream from the R2 transcription start site, and on regulatory proteins that bind to these sequences The aim of these studies is both to understand how transcription from a natural promoter is regulated in detail, but also to identify potential targets for interference with the expression of the R2 gene Inhibition of ribonucleotide reductase activity by hydroxyurea, which is a specific inhibitor of the R2 subunit, or by peptidomimetics that interfere with the interaction between the herpes simplex virus R1 and R2 subunits, has previously proven to be useful for antiproliferative therapy [28,29] In order to extend these studies to also include the function of sequences around the R2 transcription start site, we obtained a polyamide that bound specifically to the sequence immediately upstream from the mouse R2 TATA-box Polyamides are described as potent inhibitors of protein–DNA interactions and they penetrate the plasma membrane efficiently [30] Our initial experiments presented here showed that the R2-specific polyamide bound efficiently to its target sequence in the R2 promoter but we found no inhibition of expression from an R2 promoter-luciferase reporter gene in stably transformed cells, even at high concentrations of polyamide We realized that we had taken for granted that the R2 TATA-box is essential for transcription initiation at the R2 promoter and had overseen the possibility that the R2 TATA-box could be redundant Our results presented here show that the polyamide could compete with recombinant TBP for binding to the mouse R2 TATAbox and that both TBP and the R2 TATA-box are required for transcription from the R2 promoter in an in vitro transcription system reconstituted from recombinant or highly purified mouse general RNA polymerase II transcription factors In these assays we used naked plasmid DNA templates where the G-less cassette is fused directly downstream from the mapped transcription start of the full-length R2 promoter, which either contained the normal R2 TATA-box, or a mutation of the TATA-box to GCGCGC However, these results not reflect the situation in vivo, as our reconstituted in vitro transcription system lacks both TAFs and coactivators like mediator Initially, we therefore made corresponding R2-promoter constructs fused to the luciferase reporter gene to study the importance of the R2 TATA-box in vivo Similar to the experiments using the polyamide in vitro, that Table Comparison of the relative transcription levels from the different R2-promoter gene constructs used in vivo and in vitro Relative values in vivo TATAWt-UTRWt TATAmut-UTRWt TATAWt-UTRmut TATAmut-UTRmut Relative values in vitro 0.7 2.90 1.03 0.73 6.14 1.46 indicated redundancy of the R2 TATA-box, we found no requirement for the R2 TATA-box in vivo We had first neglected that the two types of templates used here, the R2-promoter coupled to either the G-less cassette or to the luciferase reporter gene, differed as the luciferase constructs also contained 17 base pairs of the mouse R2 5¢-UTR We therefore made new reporter gene constructs, both G-less cassette constructs including the R2 5¢-UTR, and luciferase constructs where the R2 5¢-UTR was replaced by sequences from the G-less cassette and the luciferase 5¢-UTR For each construct we also made a corresponding version with a mutated R2 TATA-box By comparison of all constructs we could now find a common theme for the dependency of the R2 TATA-box (Table 1) Both in vivo and in vitro, a mutation of the R2 TATA-box had a very limited effect on transcription from the R2 promoter in the presence of the R2 5¢-UTR In contrast, the R2-TATA box was required for full expression from R2-promoter templates lacking the 5¢-UTR However, the requirement of the R2 TATA-box was not absolute in the later situation Rather, either the lack of a 5¢-UTR (G-less cassette controlled by the TATAWtUTRmut R2 promoter) or a mutation of the R2 5¢-UTR (luciferase reporter controlled by the TATAWtUTRmut promoter) results in an up-regulation of transcription In this background, an additional mutation of the R2 TATA-box brings transcription down to the levels observed for the constructs that contain the R2 5¢-UTR (G-less cassette and luciferase under the control of the TATAWtUTRWt or TATAmutUTRWt promoters) The results presented above led us to focus on the R2 5¢-UTR Sequences downstream from the transcription start site have previously been identified as essential for transcription from different promoters, especially from those lacking a TATA-box The most well studied example is the downstream promoter element (DPE) which was identified as a sequence present in many TATA-less promoters, and which interacts with the TFIID-complex [14] However, more recent studies showed that DPE interacts with the Drosophila homolog of the transcriptional repressor known as NC2 or Dr1-Drap1 and that purified recombinant dNC2 activates DPE-containing promoters and represses TATA-containing promoters [31] Detailed studies of protein–DNA interactions between the human TFIID complex and the adenovirus major late promoter using cross-linking have shown that TFIID subunits contact DNA both downstream and upstream of the TATA-box, but also that the hTAFII135 and hTAFII250 interact with sequences even downstream of the transcription start site Several recent reports also show that TFIID or specific TAF subunits are involved in repression of transcription rather Ó FEBS 2003 1800 I Kotova et al (Eur J Biochem 270) than activation For example, human TAFII130 interacts with heterochromatin protein (HP1) to mediate transcriptional repression, and TAFII250 has been shown both to bind to the DNA-binding domain of TBP to inhibit TBP:DNA interactions and to be involved in repression of MHC class I expression by the HIV protein Tat [32–34] While these results would fit with the regulation of the mouse R2 promoter as we have presented it here, we were unable to find a consensus DPE-sequence [(A/G)G(A/T) CGTG] in the mouse R2 5¢-UTR Instead, comparison of the mouse and human R2 promoters from the TATA-box to 30 base pairs downstream from the R2 transcription start showed that this region is highly conserved from mouse to human In this region 39 out of 59 basepairs (66%) are identical, without including any gaps, between the mouse and human sequences (Fig 5A) This is in contrast to the sequences immediately upstream or downstream from this region, which shows