Transcription by Methanothermobacter thermautotrophicus RNA Polymerase In Vitro Releases Archaeal Transcription Factor B but Not TATA-Box Binding Protein from the Template DNA

Abstract

Transcription initiation in Archaea requires the assembly of a preinitiation complex containing the TATA- box binding protein (TBP), transcription factor B (TFB), and RNA polymerase (RNAP). The results reported establish the fate of Methanothermobacter thermautotrophicus TBP and TFB following transcription initiation by M. thermautotrophicus RNAP in vitro. TFB is released after initiation, during extension of the transcript from 4 to 24 nucleotides, but TBP remains bound to the template DNA. Regulation of archaeal transcription initiation by a repressor competition with TBP for TATA-box region binding must accommodate this observation.

Archaeal RNA polymerases (RNAPs) are closely related to eukaryotic RNAP II (11, 12, 24, 27, 28). As this predicts, archaeal RNAPs do not bind directly to promoter DNA but rather are recruited to the site of transcription initiation by complex formation with archaeal homologs of the eukaryotic general transcription factors, TATA-box binding protein (TBP) and transcription factor IIB (designated TFB in Archaea [6, 21, 24]). Within a promoter, archaeal TBP binds to a TATA-box sequence located ∼25 bp upstream of the site of transcription initiation, and the strength of this interaction is a direct determinant of promoter activity (19, 24). TFB binds to a purine-rich TFB-responsive element (BRE) upstream of the TATA box and also downstream of the TBP/TATA complex to DNA near the site of transcription initiation (4, 22). Contacts between TFB and RNAP primarily position and orient the RNAP appropriately for transcription initiation (7), and robust promoter-directed transcription initiation occurs in vitro in reaction mixtures that contain only template DNA, TBP, TFB, RNAP, and nucleoside triphosphates (NTPs). Regulatory studies have revealed that archaeal transcription initiation is inhibited by repressors that bind to sequences that overlap the BRE/TATA box or the site of transcription initiation. This apparently sterically prevents TFB/TBP access to the BRE/TATA box or RNAP access to the site of transcription initiation (6, 21). For the autorepressor Lrs14 (5), Lrs14 and TBP/TFB binding to the BRE-TATA-box region upstream of lrs14 have been shown to be mutually exclusive, with added Lrs14 incapable of dislodging prebound TFB/TBP in vitro and vice versa. In the one archaeal transcription activation system reproduced in vitro, the activator binds upstream of the BRE/TATA box, and this increases the affinity of TBP for the TATA-box sequence (19).

Consistent with the archaeal regulatory systems characterized to date superficially resembling bacterial systems, archaeal and bacterial regulators do appear to have common ancestries (1, 16, 20). Repressors apparently block, and activators stimulate, transcription initiation by binding to DNA sequences in upstream regions (6, 14, 21, 24). But, the regulated events are TBP/TFB binding to the BRE/TATA-box region or RNAP recruitment to the site of transcription initiation, as opposed to bacterial sigma factor-directed binding of bacterial RNAP to promoter DNA. Regulation by binding to a specific sequence that overlaps the site of transcription initiation seems conceptually straightforward, but regulation by repressor competition with TFB/TBP for the TATA-box sequence requires more consideration. Most Bacteria have multiple sigma factors that bind to different sequences in different promoters, and although some halophilic Archaea have several TBPs and/or TFBs (3), most Archaea have only one TBP and/or TFB. These proteins must presumably therefore recognize and bind to very similar BRE/TATA-box sequences upstream of many genes. To incorporate promoter specificity into a binding competition with TFB/TBP in these Archaea, repressors must, and apparently do also, bind to promoter-specific sequences that flank the BRE/TATA box (5, 6, 21). Of more concern, bacterial transcription initiation and RNAP departure from the promoter are accompanied by release of the initiating sigma factor, and this reestablishes any repressor-versus-RNAP holoenzyme binding competition for a promoter sequence. But, in the eukaryotic RNAP II system, TBP forms a stable complex that remains associated with the promoter DNA following transcription initiation that facilitates reinitiation (9, 13, 26, 31, 32) and, based on the homology of the archaeal and RNAP II systems, archaeal TBP should remain bound to the promoter after transcription initiation. Under such circumstances, any repressor-versus-TBP binding competition could be limited to the first round of transcription. In eukaryotes, MOT1, an essential protein in yeast (2), catalyzes the ATP-dependent dissociation of TBP from promoter DNA, but archaeal genome sequences do not encode any clearly recognizable relatives of MOT1. Experiments have therefore been undertaken explicitly to determine the fate of the archaeal TBP and TFB following transcription initiation. The results obtained confirm that in an in vitro transcription system derived from Methanothermobacter thermautotrophicus (12, 29), archaeal TBP does remain associated with the template DNA after transcription initiation, whereas TFB is released shortly after initiation.

Templates and in vitro transcription.

Construction of template A (TA; 284 bp), isolation and purification of native M. thermautotrophicus RNAP, recombinant His₆-TFB and His₆-TBP, the assembly of in vitro transcription reaction mixtures, and the characterization of [³²P]CTP-labeled transcripts synthesized in vitro after separation by denaturing polyacrylamide gel electrophoresis have been described previously in detail (12, 29, 30). The relevant features of TA, and of template B (TB; 225 bp), template C (TC; 475 bp), and template D (TD; 439 bp), constructed for this investigation are shown in Fig. 1A. TA has the BRE/TATA-box region from the promoter of the M. thermautotrophicus hmtB gene (MTH0254 [23]) positioned so that transcription initiates at the start of a 24-bp sequence, a U-less cassette, that does not require UTP for transcription (29). In the presence of only ATP, CTP, and GTP, transcription initiates but stalls after transcription of the U-less cassette, resulting in stable elongation complexes that contain the stalled 24-nucleotide (nt) U-less transcript, RNAP, and template DNA. When all four NTPs are present, or when UTP is added to stalled complexes (30), transcription of TA generates a 225-nt transcript. Heparin addition blocks archaeal transcription initiation but not elongation (4, 22), and a comparison of the transcripts synthesized from TA in the presence and absence of heparin confirmed that multiple rounds of transcription initiation and runoff transcript synthesis occur in the in vitro system used during 30 min of incubation at 58°C. Addition of UTP to stalled complexes resulted in immediate elongation of the 24-nt U-less transcripts into 225-nt runoff transcripts but, with heparin also added, there was no further increase in the amount of this transcript. In contrast, in the absence of heparin, 225-nt runoff transcripts accumulated continuously after UTP addition during a 30-min incubation at 58°C (Fig. 1B).

FIG. 1.

Transcription templates and heparin inhibition of transcription initiation. (A) Construction and the sequence of TA have been published previously (Fig. 1 in reference 30). Boxes indicate the locations of the BRE, TATA box, and U-less cassette (U⁻) and, as indicated by the arrow, transcription initiated at the start of the U-less cassette results in a 225-nt runoff transcript. The TA sequence was cloned between the NsiI (N) and HindIII sites of pLITMUS28 (New England BioLabs) and was amplified for template use from the resulting plasmid (pYX6) with a biotin molecule (o ring on left) attached to the 5′ nucleotide by using primers TD2 and MX1 with the sequences 5′-CTCAGAAAAACCTTAAAATTAGCGATATATTTATATA and 5′-CTAGAACCGGTGACGTCACCA, respectively. TB has the same sequence as TA except for a 60-bp deletion (▵) as indicated. TB was also cloned between the NsiI (N) and HindIII sites of pLITMUS28, resulting in plasmid pYX2, and was amplified using primers TD2 and MX1. TC was amplified from pYX2 using primers MX1 and MX2 (5′-GTCAGGGGGGCGGAGCCTATGG), and TD was amplified from pLITMUS28 DNA using primers MX2 and MX3 (5′-CGCCAGGGTTTTCCCAGTCACGACGTT). The SpeI site (S) used in this study is indicated. (B) TA (200 ng) was incubated with TBP (100 ng), TFB (600 ng), RNAP (10 μl), 400 μM ATP, 400 μM GTP, 20 μM CTP, and 10 μCi of [α-³²P]CTP (3 kCi/mmol; ICN, Costa Mesa, Calif.) in 100 μl of transcription buffer for 8 min at 58°C to allow the formation of stalled elongation complexes. The reaction mixture was divided in half, and 1 μl of heparin (4 mg/ml; U.S. Biochemicals, Cleveland, Ohio) was added to one of the two resulting reaction mixtures. UTP (400 μM, final concentration) was then added to both, and aliquots (8 μl) were removed after 2, 5, 10, 15, 20, and 30 min of incubation at 58°C and mixed with an equal volume of 95% formamide-20 mM EDTA-0.05% bromophenol blue-0.05% xylene cyanol. The transcripts present were separated by electrophoresis through a 6% denaturing polyacrylamide gel, visualized, and quantified by phosphorimaging (Storm model 840; Amersham Biosciences) as previously described (12, 29, 30). The phosphorimage shown demonstrates the relative amounts of the 225-nt transcript synthesized in the presence (○) and absence (•) of heparin quantified in the graph.

TB has the same sequence as TA, except for a 60-bp deletion of DNA located immediately downstream of the U-less cassette in TA (Fig. 1A). TB transcription in the absence of UTP therefore also generated stable stalled elongation complexes, and transcription with all four NTPs present resulted in 165-nt runoff transcripts. TC has the entire TB sequence fused downstream from a 250-bp sequence derived from pLITMUS28 (nt 2229 to 2479; New England BioLabs, Beverly, Mass.). TD was just pLITMUS28 DNA, specifically the sequence from nt 2229 to 2668 (Fig. 1A). PCR amplification was used to generate preparative amounts of the template DNAs, and one of the primers in each reaction mixture had a biotin molecule attached to the 5′ nucleotide (Fig. 1A). These template DNAs, plus any proteins bound to the DNA, could therefore be removed from a reaction mixture, by magnetic attraction, after affinity binding to streptavidin-coated magnetic beads (DynaBeads; Dynal Biotech, Lake Success, N.Y.). They could then be washed and, as needed, added to second and third reaction mixtures.

Template competition assays.

Template competition assays (32) were used to determine if transcription released TBP and/or TFB from TA and TB. TA DNA (100 ng; upper experiment in Fig. 2A) or TB DNA (100 ng; lower experiment in Fig. 2A) was incubated with TBP (50 ng), TFB (300 ng), and RNAP (5 μl) in transcription buffer (120 mM KCl, 10 mM MgCl₂, 2 mM dithiothreitol, 20 mM Tris · HCl [pH 8]) for 20 min at 20°C, magnetic beads (10 μg) were added, and incubation continued at 20°C for 5 min. The beads and attached complexes were removed, washed four times with 200 μl of transcription buffer, and mixed with 80 ng of TB (Fig. 2A, upper experiment) or TA (Fig. 2A, lower experiment), TFB (300 ng), and RNAP (5 μl) in 50 μl of transcription buffer. After incubation for 10 min at 20°C, ATP, UTP, GTP (each at a 400 μM final concentration), and CTP (20 μM; 10 μCi of [α-³²P]CTP) were added, and the reaction mixtures were placed at 58°C. Aliquots were taken after 10, 20, and 30 min, and the ³²P-labeled transcripts present (225 nt from TA, 165 nt from TB) were separated by polyacrylamide gel electrophoresis and visualized by phosphorimaging (Fig. 2A). The logic of the experiment is that, after NTP addition, transcription will occur and generate a runoff transcript from the template incubated initially with RNAP, TBP, and TFB, and if this releases TBP, that protein will be available to facilitate transcription initiation from the second template (32). If transcription of the first template does not release TBP, then the second template will not be transcribed unless TBP is added, as a supplement, to the reaction mixture. As shown in Fig. 2A, only transcripts from the template preincubated with TBP were synthesized in the reaction mixtures that contained both templates but no TBP supplement. In contrast, when the same procedure was used to assay for TFB release, transcripts were synthesized from both templates in mixtures of the templates without a TFB supplement (Fig. 2B). Essentially the same result, leading to the same conclusion, was reached in the analogous experiments undertaken using eukaryotic RNAP II, TBP, and TFIIB (31, 32), namely, that TBP remains bound but that TFB (TFIIB) is released from the template DNA by transcription.

FIG. 2.

Template competition assays with reaction components, steps, and sampling times diagrammed. (A) TA (upper experiment) or TB (lower experiment) DNA was incubated with TBP, TFB, and RNAP, and the complexes formed were removed, washed, and added to the second reaction mixture that lacked TBP. After 10 min at 20°C, NTPs (with [³²P]CTP) were added, the reaction mixtures were placed at 58°C, and the ³²P-labeled transcripts present in aliquots taken after 10, 20, and 30 min of incubation were separated by PAGE (upper experiment, lanes 4, 5, and 6; lower experiment, lanes 7, 8, and 9) and visualized by phosphorimaging. As illustrated for the upper experiment, TBP (50 ng) was added in control reactions and the transcripts present 10, 20, and 30 min after NTP addition were separated in lanes 1, 2, and 3. Lane S contained size standards. (B) The same experiments as described for panel A, except that the complexes formed initially were subsequently mixed and incubated with TB DNA (upper experiment) or TA DNA (lower experiment) plus RNAP and TBP and not TFB. For the controls, TFB (300 ng) was added after mixing the two templates but before NTP addition.

Transcript elongation from 4 to 24 nt releases TFB.

TC and TD were constructed to provide a second approach to determining the fates of TBP and TFB after transcription initiation and also to establish when TFB was released from the template DNA. TC or TD DNA (200 ng) was incubated with TBP (50 ng), TFB (300 ng), and RNAP (5 μl) in 50 μl of transcription buffer for 10 min at 20°C. Magnetic beads (10 μg) were added, and after incubation for 10 min at 20°C the beads were removed, washed four times with transcription buffer (200 μl), and then either subjected immediately to SpeI digestion (10 U [Invitrogen, Carlsbad, Calif.] in 16 μl of transcription buffer) for 30 min at 37°C (Fig. 3A) or were first incubated under in vitro transcription conditions for 10 min at 58°C with no NTPs, ATP, ATP plus CTP, ATP plus CTP plus GTP, or all four NTPs (50-μl reaction volume; 400 μM final NTP concentrations and 10 μg of salmon sperm DNA [Sigma, St. Louis, Mo.]) before washing and SpeI digestion (Fig. 3B). After incubation with SpeI, the beads were removed, 4 μl of 5× gel loading buffer (10% sodium dodecyl sulfate [SDS], 2.5% β-mercaptoethanol, 0.05% bromophenol blue, 300 mM Tris · HCl [pH 6.8]) was added to the remaining supernatant, and the mixture was placed at 100°C for 5 min. The proteins present were then separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE; 15% polyacrylamide gel) and transferred to a Hybond nitrocellulose membrane (Amersham). The presence of His₆-TBP and/or His₆-TFB was determined by immunoblotting using anti-Xpress antibody (Invitrogen), a peroxidase-labeled anti-mouse secondary antibody, and the ECL Plus detection system (Amersham). TC and TD have one SpeI site (Fig. 1A), and the template DNA distal to this site, relative to the bead, plus any proteins bound to this DNA were released from the beads by SpeI digestion and remained in the supernatant after bead removal. When the control TD was used that has no BRE/TATA-box-related sequences, no detectable TFB or TBP remained in the supernatant after SpeI digestion and bead removal (Fig. 3A). However, both transcription factors were present after bead removal when TC was used (Fig. 3A), consistent with TFB and TBP binding to the BRE/TATA box located distal to the SpeI site in TC (Fig. 1A). Both transcription factors were also present in the supernatant after bead removal when the TC complexes were incubated in transcription reaction mixtures that contained ATP or ATP plus CTP before SpeI digestion. But, after incubation in transcription reaction mixtures that contained ATP plus CTP plus GTP, or all four NTPs, only TBP remained in the supernatant. The sequence of the U-less transcript begins 5′-ACACGGA (Fig. 1A), and transcription is therefore initiated and the first 4 bp of the template are transcribed in reaction mixtures that contain only ATP plus CTP. Under such circumstances, TFB and TBP remained attached to the TC DNA located distal to the SpeI site but, with GTP also present, allowing transcription to extend to the end of the 24-bp U-less cassette, TFB but not TBP was released from the template DNA (Fig. 3B).

FIG. 3.

Immunoblot assays of template binding by TFB and TBP. (A) TC or TD was incubated with TBP, TFB, and RNAP in transcription buffer for 10 min at 20°C. The complexes formed were removed, washed, and subjected to SpeI digestion, and the transcription factors remaining in the supernatant after bead removal were separated by SDS-PAGE and detected by immunoblotting. (B) After washing, aliquots of TC-containing complexes were incubated under transcription conditions in reaction mixtures that contained either no NTPs (−), ATP (A), ATP plus CTP (AC), ATP plus CTP plus GTP (ACG), or all four NTPs (ACGU). The complexes were then washed and subjected to SpeI digestion. After bead removal, the presence of TFB and/or TBP in the supernatant of the SpeI digest was determined by immunoblotting. Given the 5′ sequence of the U-less cassette (Fig. 1A), transcription on TC in the presence of ATP, ATP plus CTP, and ATP plus CTP plus GTP extended to positions 1, 4, and 24, respectively. With all four NTPs present, 165-nt runoff transcripts were generated.

Discussion.

Based on homology of the archaeal and RNAP II basal transcription machineries (6, 21, 24), archaeal TBP was predicted to remain attached but TFB to dissociate from the template DNA following transcription initiation. The results obtained confirm this prediction for transcription in vitro. They are also consistent with footprinting data, which demonstrated that the TATA box region of an archaeal promoter remained protected from exoIII digestion after transcription initiation in vitro (25) and that a structural transition occurs in the elongation complex during transcript elongation from positions +7 and +9 (25), most likely coinciding with the first translocation forward of the archaeal RNAP. The RNA-DNA hybrid within such an archaeal elongation complex is 9 to 12 bp (25), and for transcription to continue beyond 12 nt, the 5′ end of the nascent transcript presumably has to exit the elongation complex. In this regard, TFIIB binding to RNAP II covers the transcript exit site (10), and protrusion of the nascent transcript has been proposed to dissociate TFIIB from RNAP II and to so facilitate promoter clearance (10, 11). The results currently available are consistent with TFB being similarly dissociated from the archaeal RNAP elongation complex by nascent transcript protrusion.

The results reported were generated in vitro, in a minimal initiation system, and additional transcription factors undoubtedly contribute in vivo to the assembly, stabilization, and longevity of archaeal initiation complexes. However, based on the RNAP II precedent (8), archaeal TBP may well remain bound to genomic DNA in vivo until removed, possibly by an unrecognized archaeal analog of MOT1 or as a consequence of more generic events, such as DNA replication, transcription from an upstream promoter (17), or DNA distortion by chromatin-forming proteins (21). Proteins with TBP-binding activity have been identified in Pyrococcus species (18, 24), although they are not widely conserved in Archaea, but all Archaea do have DNA-distorting chromatin proteins. In considering this, NC2 is a very widely conserved eukaryotic regulator that interacts specifically with the DNA in TBP-TATA complexes by employing a histone fold-based mechanism of DNA binding (8, 15), and the archaeal histones present in almost all Euryarchaea have very similar histone folds and DNA binding properties (21). TBP could be dislodged from promoter DNA in these Archaea by adjacent archaeal histone binding, DNA compaction, and distortion. Clearly, this hypothesis cannot be extended directly to the Crenarchaea, which lack histones, but the underlying concept remains valid as these species do contain substantial amounts of nonhistone protein but nevertheless highly DNA-distorting chromatin proteins (14, 21).