INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

In prokaryotes and eukaryotes, transcription initiation can be divided into three basic steps: assembly of a closed initiation complex at the promoter, isomerization of the closed complex to the open complex, and promoter clearance (4, 11, 19, 20, 43). In principle, transcriptional regulators can affect any of these steps. For example, the Escherichia coli protein CAP (catabolite activator protein) has been shown to facilitate the binding of RNA polymerase to the promoter, isomerization, and promoter escape (4, 10, 30, 41, 43). In eukaryotes, the transactivation domain of the herpes simplex virus protein VP16 has been shown to stimulate transcription initiation, perhaps by interacting with TFIIB (18, 37), TFIIH (60), and TFIID (29). Therefore, VP16 may have a role in promoter assembly. Yankulov et al. (66) have demonstrated that the VP16 transactivation domain may also stimulate elongation, possibly by increasing the processivity of RNA polymerase II. Other activators, like the human immunodeficiency virus TAT protein, may affect still other steps (26).

Careful order-of-addition experiments with purified components of the general transcription machinery have suggested a stepwise assembly of initiation complexes in vitro. According to this model, the TATA box (or another core promoter element) is first recognized by TBP, the TATA box-binding subunit of the TFIID complex. TFIIA and TFIIB then join promoter-bound TFIID. The resulting TFIID-TFIIA-TFIIB (DAB)-promoter complex subsequently recruits RNA polymerase II and TFIIF. Finally, TFIIE and TFIIH enter the initiation complex, and isomerization can occur (2, 3, 8, 44, 45, 51, 53). A somewhat different view of initiation complex assembly has emerged with the discovery of a large multisubunit RNA polymerase II complex in yeast cells; this complex is called the holoenzyme (for reviews, see references 22, 23, 32, and 68). Such yeast holoenzyme complexes have been reported, depending on the method of isolation and analysis, to contain RNA polymerase II; SRB proteins; TFIIF, TFIIB, and TFIIH (28, 31); Sin4P, Rgr1P, and Gal11P (34); and polypeptides of the SWI-SNF complex (65). RNA polymerase II holoenzyme complexes have recently also been isolated from mammalian cells (5, 7, 39, 47, 48, 54). In three cases, such complexes have been enriched by a single affinity purification step with an immobilized CDK7 antibody (47); the immobilized elongation factors, elongin A or TF-IIS (48); or an immobilized TFIIF antibody (7). In two of these cases (47, 48), all general transcription factors required for promoter-specific initiation could be recovered. Quantitative immunoblot experiments by Pan et al. (48) revealed nearly stoichiometric amounts of the largest RNA polymerase II subunit RPB1 and TFIIB, TFIID, TFIIE, TFIIF, and TFIIH in the affinity-purified holoenzyme complex. Since all of these polypeptides coeluted in gel filtration analyses, it is likely that they are part of a large complex with a molecular mass of about 2 × 10⁶ Da. Recently, holoenzyme complexes capable of autonomous transcription initiation have also been described for RNA polymerase I (52, 55) and RNA polymerase III (62). Evidence for the association of RNA polymerase III with its two essential initiation factors, TFIIIB and TFIIIC, in the absence of DNA had already been presented more than 10 years ago by Wingender et al. (64).

The discovery of the RNA polymerase II holoenzyme has substantially modified our view of initiation complex assembly and the way sequence-specific transcription factors participate in this process. At least at some promoters, initiation complex assembly could occur in a single step, in a way similar to the binding of bacterial RNA polymerase holoenzyme to promoters. As a consequence, transactivators may stimulate transcription simply by assisting in the recruitment of RNA polymerase II holoenzyme (for reviews, see references 22, 23, 32, and 49). In fact, genetic studies by Ptashne and coworkers demonstrated that a fortuitous contact between a promoter bound factor and a component of the RNA polymerase II holoenzyme is sufficient to efficiently activate transcription (1, 13, 16). Thus, the DNA-binding/dimerization domain of Gal4, normally inactive in transcription activation, can efficiently stimulate transcription in yeast strains expressing Gal11P, a mutant form of the holoenzyme component Gal11. Moreover, the acidic activation domain of the viral protein VP16 interacts with the RNA polymerase II holoenzyme (24). Studies by Thompson and Young (59) suggest that the yeast holoenzyme may actually be the major form of RNA polymerase II capable of transcription initiation in vivo. The yeast cup1 gene seems to be an exception, since it can be efficiently transcribed with an RNA polymerase II enzyme lacking the CTD (40).

Experimental evidence allowing the discrimination between a sequential and a single-step assembly of the RNA polymerase II initiation complex in the living cell is difficult to obtain. We therefore resorted to in vitro transcription studies to shed more light on the initiation complex assembly. In particular, we wanted to examine whether the biochemical properties of the RNA polymerase II transcription machinery is compatible with a single-step assembly mechanism. Here we show that the basic leucine zipper protein thyroid embryonic factor (TEF), a potent transactivator both in vivo (15) and in vitro (the present study), also efficiently stimulates in vitro transcription by an immunopurified RNA polymerase II holoenzyme. Moreover, the activation domain of this transcription factor binds to an RNA polymerase II complex containing all general transcription factors required for promoter-specific transcription initiation. These findings are compatible with a mechanism by which promoter-bound transcriptional activators recruit the RNA polymerase II transcription machinery in a single step, as in models proposed for prokaryotic organisms.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Nuclear extract preparation. Nuclear-transcription-competent extracts were prepared as described by Ossipow et al. (47). Briefly, highly purified rat liver nuclei were isolated as described by Lichtsteiner et al. (36). The nuclear pellet was resuspended at an optical density at 260 nm of 1 per ml in nuclear lysis buffer containing 10 mM HEPES (pH 7.6), 100 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 0.1 mM NaF, 0.1 mM Na₃VO₄, 0.1 mM EDTA, 1 mM dithiothreitol (DTT), 0.1 mM phenylmethylsulfonyl fluoride, and 10% glycerol. KCl was added to a final concentration of 0.3 M to extract soluble nonhistone proteins. After 20 min on ice, the nuclear lysate was spun at 36,000 rpm in a fixed-angle Ti60 rotor for 60 min at 0°C to sediment the insoluble chromatin components. Soluble proteins were precipitated by the addition of 0.3 g of solid (NH₄)₂SO₄ per ml of supernatant. After 1 h on ice, the ammonium sulfate precipitate was recovered by sedimentation at 36,000 rpm in a fixed-angle Ti60 rotor for 30 min at 0°C. The precipitated proteins were resuspended in dialysis buffer (25 mM HEPES (pH 7.6), 100 mM KCl, 0.1 mM NaF, 0.1 mM Na₃VO₄, 0.1 mM EDTA, 0.25 mM DTT, and 10% glycerol) in 5% of the original nuclear lysate volume and dialyzed twice for 2 h against 250 volumes of the same buffer. After a 2-min centrifugation in an Eppendorf microfuge, the supernatant was divided into 100-µl aliquots, snap frozen, and kept under liquid nitrogen until use. The protein concentration was determined as described in Gorski et al. (21). Typically, 1 g of rat liver (wet weight) yields 100 µl of nuclear extract containing 8 to 12 mg of protein/ml.

RNA polymerase II holoenzyme immunopurification. RNA polymerase II holoenzyme was immunopurified as described in Ossipow et al. (47). Briefly, per assay, 0.5 µl of ascites fluid containing a monoclonal antibody directed against human CDK7/MO15 (MO-1.1 [56]) or an irrelevant control antibody (47) was incubated with 15 µl of a suspension containing magnetic beads coated with goat immunoglobulin G (IgG) anti-mouse IgG (Dynal M450) for 4 h at 4°C. The beads were extensively washed in phosphate-buffered saline containing 0.1% Triton X-100 and incubated with 50 µl (250 to 500 µg of protein) of undiluted crude nuclear extracts for 1 h at 4°C with gentle shaking. The beads were washed three times with 400 µl of washing buffer (25 mM HEPES, pH 7.6; 50 mM KCl, 0.1 mM EDTA; 0.1 mM Na₃VO₄; 0.1 mM NaF; 10% glycerol; 0.1% Triton X-100).

Preparation of magnetic beads with DNA-bound TEF proteins. First, 400 pmol of the biotinylated top strand of a 25-mer oligonucleotide encompassing the high-affinity site for TEF (and other PAR leucine zipper proteins [12]) was annealed to 400 pmol of its complementary bottom strand. The sequence of the top strand is 5'-GTTCTTGGTTACGTAATCTCCAATGGTTCTT-3' (the TEF binding site underlined). The double-stranded oligonucleotide was bound to 200 µl (2 mg) of streptavidin-coated magnetic beads (Dynal M280) according to the manufacturer's specifications. The decorated beads were incubated for 45 min at 4°C with 100 µg of either full-length TEF or TEF lacking the activation domain in a buffer containing 10 mM Tris (pH 7.4), 0.1 mM EDTA, 200 mM NaCl, 10 mM DTT, and 0.1% Triton X-100. The beads were washed and stored in the same buffer.

Preparation of magnetic beads with covalently coupled GST fusion proteins. The fusion proteins glutathione S-transferase (GST)-DBP (amino acids 127 to 208) and GST-TEF (amino acids 14 to 159) were overexpressed in E. coli (strain BL21), purified, and eluted from the glutathione beads according to the manufacturer's specifications (Pharmacia). The purified proteins were precipitated by the addition of 0.3 g of solid (NH₄)₂SO₄ per ml and then resuspended in a buffer containing 100 mM potassium phosphate (pH 7.8) and 10 mM DTT. They were then loaded onto a 10-ml G50 size exclusion column, and the protein-containing fractions were pooled. The proteins were concentrated by centrifugation in a Centricon 30 tube to a final concentration of 8 mg/ml. Then, 1 mg of each fusion protein was covalently coupled to 166 µl (5 mg) of MPG glyceryl-porous magnetic beads (MGLY0502; CPG Inc., Lincoln Park, N.J.) in a buffer containing 100 mM potassium phosphate (pH 7.8) according to the manufacturer's indications. The protein concentration on the beads, which was determined by including a trace of radiolabeled fusion protein in the coupling reaction, was estimated to be 120 µg/mg of beads.

Affinity enrichment of RNA polymerase II holoenzyme by adsorption to the TEF activation domain. For each reaction, 60 µg of covalently coupled GST fusion proteins or 200 ng of full-length and truncated TEF bound to DNA-coated magnetic beads were used. The beads were incubated with 50 µl (ca. 500 µg of protein) of crude nuclear proteins for 20 min at 4°C with gentle shaking. They were then gently washed three times with 200 µl of washing buffer (25 mM HEPES, pH 7.6; 50 mM KCl; 0.1 mM EDTA; 0.1 mM Na₃VO₄; 0.1 mM NaF; 10% glycerol; 0.1% Triton X-100) and directly used for in vitro transcription.

In vitro transcription. In vitro transcription reactions were performed as described by Gorski et al. (21) with either crude nuclear extract (10-µl assays) or immunopurified and affinity-purified RNA polymerase II holoenzymes immobilized on washed beads (10- to 20-µl assays). The reactions typically contained 1 µg of template DNA/10 µl and were incubated for 45 min at 37°C with gentle shaking.

Size fractionation of nuclear proteins by gel filtration. First, 7 mg (700 µl) of nuclear extract was incubated for 20 min at 4°C with 5 mM MgCl₂, 10 µM distamycin, 20 µg of ethidium bromide per ml, and 100 µg of RNase A per ml. The extract was spun at 4°C for 15 min at 15,000 rpm to remove the insoluble material, and the supernatant was loaded on a 25-ml Sepharose CL2B column. The chromatography was performed at 4°C, with a flow rate of 500 µl/min, in a buffer containing 25 mM HEPES (pH 7.6), 100 mM KCl, 5 mM MgCl₂, 0.1 mM NaF, 0.1 mM Na₃VO₄, 0.1 mM EDTA, 0.25 mM DTT, and 10% glycerol, and 500-µl fractions were collected. For RPB1, TFIIB, TFIIE, and TBP, 20 µl of crude fractions were used for Western blot analysis. For TFIIE, TFIIH, CDK7, and SRB7, 100 µl of the fractions were trichloroacetic acid precipitated before Western blot analysis. Then 15 µl of the crude fractions was directly used in an in vitro transcription reaction, or 100 µl was concentrated sevenfold by centrifugation in a Centricon 30 tube before immunopurification with the CDK7 antibody and in vitro transcription. Western blotting was performed as described in Ossipow et al. (46, 47).

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Recombinant TEF activates in vitro transcription by immunopurified RNA polymerase II holoenzyme. We have previously reported the isolation of an RNA polymerase II holoenzyme from rat liver nuclei by using coimmunopurification with a monoclonal antibody directed against CDK7 (MO15), the CTD kinase subunit of TFIIH (47). Our first attempts to activate transcription by this form of polymerase II with bacterially expressed C/EBP failed (47). However, even in unfractionated liver nuclear extracts, this transactivator stimulates transcription only moderately. We therefore searched for a more potent transcription factor and found that the leucine zipper transcription factor TEF (9) stimulates in vitro transcription very efficiently in liver nuclear extracts. In the experiment shown in Fig. 1, a promoter bearing nine TEF binding sites (D9Alb400 [38]) is strongly activated upon addition of E. coli-derived recombinant TEF. As expected, a control promoter lacking such binding sites, the adenovirus major late promoter fused to a 200-bp G-free cassette (AdML200), remains unaffected by TEF (Fig. 1, lane TEF). This activation is dependent on the TEF activation domain, since a truncated version containing only the DNA-binding domain does not stimulate transcription (Fig. 1, lane TEFN). Recombinant TEF also activates in vitro transcription efficiently from the albumin promoter, which contains only one TEF recognition sequence (data not shown).

View larger version (40K): [in this window] [in a new window]   FIG. 1.   Recombinant TEF is a potent transactivator of in vitro transcription in liver nuclear extracts. Two G-free cassette templates containing either the adenovirus major late promoter (AdML200, 400 to +10 [38]) or a synthetic promoter containing nine binding sites for TEF in front of the albumin core promoter (D9Alb400 [38]) were incubated with 46 µg of nuclear extract for in vitro transcription. The final concentrations of wild-type (lanes TEF) or amino-terminally truncated TEF (lanes TEFN) in the reactions are indicated above each lane.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Analysis of immunopurified RNA polymerase II holoenzyme. (A) In vitro transcription of a template containing the adenovirus major late promoter (AdML200, 400 to +10) with 50 µg of liver nuclear extract (NE) (lane 1), proteins immunoenriched with an irrelevant antibody from 250 µg of NE (see reference 47) (lane 2), and proteins immunoenriched with CDK7 antibody from 250 µg of NE (lane 3). (B to E) Western blot analysis of immunopurified RNA polymerase II holoenzyme with antibodies against different nuclear proteins. In each experiment, 50 µg of liver nuclear proteins (lanes 1) were size fractionated by SDS-PAGE alongside proteins immunoenriched from 250 µg of liver nuclear extract with either the CDK7 antibody (lanes 3) or an antibody against an irrelevant peptide (lanes 2). The epitopes for the various antibodies used in these immunoblot experiments are indicated at the righthand side of the panels.

View larger version (45K): [in this window] [in a new window]   FIG. 3.   TEF activates transcription by immunopurified RNA polymerase II holoenzyme. Two G-free cassette templates containing either the adenovirus major late promoter (AdML200, 400 to +10) or a synthetic promoter composed of nine binding sites for TEF in front of the albumin core promoter (D9Alb400) were incubated with RNA polymerase II holoenzyme immunopurified from 500 µg of nuclear extract. The final concentration of TEF in the reactions is indicated above each lane.

View larger version (59K): [in this window] [in a new window]   FIG. 4.   Analysis of size-fractionated liver nuclear proteins. Nuclear proteins (7 mg) were treated with distamycin, ethidium bromide, and RNase A and then size fractionated on a Sepharose CL-2B column. Every fifth fraction (lanes 17 to 77) or the unfractionated nuclear extract (lane ne) was analyzed by Western blotting with the antibodies against the polypeptides indicated at the right hand side of the figure. In vitro transcription from the adenovirus major late promoter (template AdML200, 400 to +10) with crude CL2B fractions (IVT) and after immunopurification in the CL2B fractions with the CDK7 antibody (IP).

TEF squelches transcription at high concentrations. We observed that the addition of increasing amounts of recombinant TEF to a crude nuclear extract progressively stimulates in vitro transcription from a target promoter until it reaches a concentration of 40 nM (Fig. 5). At a TEF concentration of 400 nM, the relative activation factor is already lower than that observed at a 10-fold-lower concentration, and at 4 µM no significant stimulation of transcription can be seen (Fig. 5). The inhibitory effect of TEF at high concentrations is likely to reflect squelching, that is, competition between free and promoter-bound activator for target surfaces on the RNA polymerase II machinery (reviewed in reference 49). Since half-maximal inhibition of activated transcription is observed at TEF concentrations between 400 nM and 4 µM, the affinity (K_D) of the interaction between TEF and components of the RNA polymerase II machinery can be estimated to be in the low micromolar range. Remarkably, TEF squelches transcription from the D9 target promoter much more dramatically than from the adenovirus major late promoter, indicating that the association of TEF with the transcription machinery in solution either does not attenuate or only weakly attenuates transcription initiation as such. This result also suggests that the surfaces of general transcription factors interacting with TEF are not required for transcription from the adenovirus major late promoter.

View larger version (62K): [in this window] [in a new window]   FIG. 5.   TEF squelches transcription at high concentrations. Two G-free cassette templates containing either the adenovirus major late promoter (AdML200, 400 to +10) or a synthetic promoter composed of nine binding sites for TEF in front of the albumin core promoter (D9Alb400) were incubated for in vitro transcription with 50 µg of nuclear extract. The final concentration of recombinant TEF present in the assay is indicated above each lane.

The TEF activation domain binds to the RNA polymerase II holoenzyme. The specific squelching described in the previous section suggested that TEF interacts specifically with one or more components of the RNA polymerase II machinery in solution. Encouraged by this observation, we wanted to examine whether the activation domain of TEF can bind to an RNA polymerase II holoenzyme in the absence of promoter DNA. Towards this aim, we first immobilized biotinylated double-stranded oligonucleotides encompassing a high-affinity site for TEF (12) to streptavidin-coated magnetic beads. Full-length recombinant TEF, or an amino-terminally truncated version shown to be incompetent for transcription (Fig. 1), was then bound to the immobilized DNA. In control experiments, we have shown that these protein-DNA complexes have a half-life of much more than 1 h at 4°C (data not shown), indicating that TEF remains associated during the much shorter affinity purification procedure described below. Three sets of magnetic beads were then incubated with a transcription-competent liver nuclear transcription extract: (i) beads with the oligonucleotide alone, (ii) beads with the amino-terminally truncated TEF tethered to the oligonucleotide, and (iii) beads with full-length TEF bound to the oligonucleotide (Fig. 6A). Proteins with affinities to the beads were purified by magnetic attraction. After extensive washes, the affinity-enriched proteins were incubated with nucleoside triphosphates and a plasmid carrying the adenovirus major late promoter fused to a G-free cassette. No transcriptional activity could be recovered from beads decorated with the oligonucleotide alone or from beads with the oligonucleotide and the amino-terminally truncated version of TEF (Fig. 6B, lanes CO and TEFN). In contrast, incubation of nuclear proteins with beads containing the TEF binding site and full-length TEF resulted in the enrichment of an activity capable of accurate transcription initiation on the adenovirus major late promoter (Fig. 6B, lane TEF). This experiment suggests that the activation domain of TEF recruits either an RNA polymerase II holoenzyme or all individual components required for promoter-specific transcription-initiation to the beads. Although we cannot formally exclude the later possibility, the former appears more likely, in view of the evidence presented above for a large transcription-competent RNA polymerase II complex.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   TEF interacts with an RNA polymerase II complex containing all components required for promoter-specific initiation. (A) Schematic representation of the different paramagnetic streptavidin beads used in the experiments shown in panel B. Double-stranded oligonucleotides encompassing a high-affinity TEF binding site (sequence given in Materials and Methods) are depicted as ladders. Biotin groups and streptavidin molecules are represented by filled circles and rounded Y-like symbols, respectively. The C-terminal DNA-binding/dimerization domains of TEF are shown as bent cylinders embracing the DNA binding sites, and the N-terminal TEF activation domains are represented as cones (see also Fig. 7). (B) Affinity selection of an RNA polymerase II complex with recombinant TEF bound to its DNA recognition site. The paramagnetic beads used in these experiments are schematically depicted in panel A. The beads were incubated with 250 µg of nuclear extract and, after being washed and after magnetic attraction, they were incubated for in vitro transcription with a G-free cassette template containing the adenovirus major late promoter (AdML200, 400 to +10 [38]). The following beads were used (see panel A): lane CO, control beads with DNA alone; lane TEFN, TEFN-DNA beads; lane TEF, TEF-DNA beads.

View larger version (26K): [in this window] [in a new window]   FIG. 7.   The activation domain of TEF interacts in solution with an RNA polymerase II holoenzyme. (A) Schematic representation of the paramagnetic beads with covalently bound GST-TEF14-159 or GST-DBP127-208 fusion proteins that were used in the experiments shown in panels B and C. (B) In vitro transcription with affinity-enriched nuclear proteins. The beads schematically represented in panel A were incubated with 500 µg of liver nuclear proteins and washed. The recovered proteins were incubated for in vitro transcription with a G-free cassette template containing the adenovirus major late promoter. (C) Affinity-enriched proteins were analyzed by immunoblotting with an SRB7 antibody. Lanes: N.E., 50 µg of input nuclear extract proteins; GST-TEF, proteins recovered from 500 µg of nuclear proteins with GST-TEF magnetic beads (see panel A); GST-DBP, proteins recovered from 500 µg of nuclear proteins with GST-DBP magnetic beads (see panel A).

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

The results reported here demonstrate that the transcription factor TEF stimulates transcription in a simple reaction consisting of E. coli-derived recombinant TEF, a target promoter containing multiple TEF recognition sequences, and an immunopurified RNA polymerase II holoenzyme. The direct interaction of TEF with components of the RNA polymerase II machinery is supported by two additional findings. First, high concentrations of TEF result in specific squelching of target gene transcription in nuclear extracts (Fig. 5). This suggests that excess TEF that is not bound to promoter sites saturates its target surfaces on general transcription factors, thereby competing with promoter-bound TEF for the interaction with such components. Second, the activation domain of TEF, when immobilized on paramagnetic beads, can be used to affinity enrich all components of the RNA polymerase II transcription machinery required for promoter-specific initiation. Since SRB7, a hallmark of the large RNA polymerase II holoenzyme complex, is also among the peptides affinity enriched with TEF, it is likely that the general transcription factors are recruited to the paramagnetic beads in form of an RNA polymerase II holoenzyme. The TEF activation domain binds to this complex irrespective of whether it is tethered to the beads via binding of full-length TEF to immobilized DNA recognition sequences or via a TEF-GST fusion protein. Therefore, DNA binding of TEF is not required for its interactions with RNA polymerase II components.

The direct binding of an initiation-competent RNA polymerase II holoenzyme to the activation domain of TEF suggests a simple mechanism for transcription activation: TEF may recruit the complete transcription apparatus to the promoter in a single step. Obviously, the feasibility of such a mechanism depends on the presence of all essential initiation factors in the RNA polymerase II holoenzyme. A particularly important question is whether TFIID, the major promoter-binding component of the RNA polymerase II machinery, is associated with a fraction of the holoenzyme (see below). All authors who applied fast single-step affinity purifications have recovered some TFIID in their preparations. Thus, the presence of TBP could be demonstrated in RNA polymerase II holoenzymes immunopurified with antibodies against CDK7 kinase (47; this study), TFIIF (7), TFIIE (46a), and SRB5 (69). Likewise, TFIID was found to be associated with holoenzymes enriched by affinity chromatography on resins containing the elongation factors TFIIS or elongin A (48). In their report, Pan et al. (48) present careful stoichiometry measurements and find that about half of the affinity-enriched RNA polymerase II complexes contain TBP. It appears likely, therefore, that the failure of some authors to recover TFIID in their holoenzyme preparations is due to the lengthy and harsh purification procedures they used. In spite of the demonstration of a holoenzyme complex containing all essential initiation factors, one would expect that a fraction of RNA polymerase complexes are devoid of TFIID. According to the model proposed by Pan et al. (48), the first initiation event may be accomplished by an RNA polymerase II holoenzyme complex containing all essential initiation factors. After promoter escape has occurred, promoter-bound TFIID may serve as a landing pad for further initiations by RNA polymerases not associated with TFIID that have terminated transcription on the same (or other nearby) gene(s). In fact, many genes are transcribed in bursts of multiple rounds (42, 63). In these cases, TFIID may remain bound to the promoter during a time period, allowing repeated initiation events to occur.

We have noticed that TEF stimulates in vitro transcription from target templates with a higher amplitude with unfractionated nuclear extracts (up to 30-fold) compared to immunopurified holoenzyme (about 4-fold). Conceivably, some proteins, such as coactivators that contribute to the overall activation by TEF, are depleted during the immunoenrichment of the RNA polymerase II holoenzyme. Alternatively, repressors of basal transcription, such as NC1 or NC2/Dr1 (27, 67), may be present in nuclear extracts but absent from purified holoenzyme. The absence of such global repressors would result in a higher basal in vitro transcription with purified RNA polymerase II holoenzyme and therefore in a lower amplitude of activated transcription.

The identification of RNA polymerase II holoenzyme polypeptides that interact with the activation domain of TEF will be a major challenge. We estimate that the K_D for the interaction of the TEF activation domain with the RNA polymerase II transcription machinery must be in the low micromolar range (see Results section). A K_D of 10⁶ M corresponds to a free energy of about 8 kcal/mol. If this binding energy were distributed over multiple TEF contacts with different components of the RNA polymerase II holoenzyme, the interaction of TEF with individual binding partners would be too weak for conventional protein-protein interaction studies. Indeed, attempts to identify TEF-interacting polypeptides of the holoenzyme by far-Western blotting techniques have failed thus far (data not shown).

The finding of holoenzyme complexes for all three forms of eukaryotic RNA polymerases (for references, see the Introduction) indicates that transcription initiation and its stimulation by sequence-specific regulatory proteins are conceptually more similar in prokaryotes and eukaryotes than hitherto assumed. Like bacterial RNA polymerase holoenzymes, all three eukaryotic RNA polymerases may occupy at least some promoters in a single step and dissociate from promoter-specific accessory factors upon promoter escape. During or after termination, the RNA polymerase core enzyme may be reassembled into an initiation competent holoenzyme, in a way similar to the transcription cycles established for bacterial RNA polymerases and sigma factors. One obvious difference between prokaryotic and eukaryotic transcription still persists, however. Most sigma factors cannot bind promoter DNA autonomously and thus have to assemble with core RNA polymerase for each initiation event. In the case of eukaryotic transcription, only the first initiation event on a given gene may require an RNA polymerase II holoenzyme containing all initiation factors. The rapidly following subsequent rounds may then be performed with incomplete RNA polymerase II complexes on already-promoter-bound initiation factors (48). The examination of such mechanisms will require novel technologies that allow multiple rounds of in vitro transcription by RNA polymerase II on the same DNA template.