Previous Article | Next Article ![]()
Molecular and Cellular Biology, February 2004, p. 1122-1131, Vol. 24, No. 3
0270-7306/04/$08.00+0 DOI: 10.1128/MCB.24.3.1122-1131.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Laboratory of Gene Transcription, Institut de Recherches Cliniques de Montréal, Montréal, Québec, Canada H2W 1R7
Received 24 September 2003/ Accepted 3 November 2003
|
|
|---|
closely approaches promoter DNA at nucleotide -10, a position immediately upstream of the transcription bubble in the open complex. Increased stimulation of transcription initiation by TFIIE
is obtained when the DNA template is artificially premelted in the -11/-1 region, suggesting that TFIIE
facilitates open complex formation, possibly through its interaction with the upstream end of the partially opened transcription bubble. These results support the central roles of the mobile clamp of RNAP II and TFIIE in transcription initiation. |
|
|---|
The availability of crystal structures for both eukaryotic (6, 7, 14, 15) and prokaryotic (4, 33, 34, 55, 59) RNAPs has been invaluable for the understanding of the many molecular features of the transcription reaction. For example, the structure of elongating RNAP II has revealed the position of the RNA-DNA duplex formed during the transcription reaction (15). The available structures support a model in which the DNA enters the enzyme through a channel formed by a pair of "jaws" before accessing a deep cleft, at the bottom of which is buried the active site with its Mg2+ ions; the DNA then turns by
90o along a wall, where the upstream end exits the enzyme (15). Recently, the crystal structure of the Thermus aquaticus RNAP holoenzyme bound to a fork junction DNA fragment was resolved (33). The structure shows that the DNA lies across one face of the holoenzyme, completely outside the active-site channel. A similar high-resolution structure remains to be obtained for RNAP II.
Over the years, site-specific protein-DNA photo-cross-linking has been used to analyze the topological organization of RNAP II complexes bound to DNA. The results indicate that the RNAP II preinitiation complex forms a very compact structure in which the DNA is tightly bent and wrapped against the core of the complex (13, 45). Our proposed model, named the DNA-wrapping model, describes basal transcription mechanisms (5). However, two puzzling features stand out from our cross-linking results: (i) some subunits of the complex cross-link to a very extensive region of the promoter, and (ii) some promoter positions can cross-link many different polypeptides. One possible explanation for these observations stems from the fact that our cross-linking procedure was performed on protein-DNA complexes that were not purified after assembly, possibly allowing the formation of a heterogeneous subpopulation of minor complexes in addition to the main complex. These subcomplexes may result from nonspecific interactions between components of the transcription machinery and promoter DNA. In order to circumvent this possible caveat and to improve the resolution of our photo-cross-linking procedure, we have developed an in-gel protein-DNA photo-cross-linking method in which protein-DNA complexes are separated using native gel electrophoresis prior to their UV irradiation and extraction from the gel. We report here the use of this method to study cross-linking of the transcription machinery assembled into a TBP-TFIIB-TFIIF-RNAP II-TFIIE-promoter complex. As expected, this in-gel photo-cross-linking methodology increases the resolution by allowing more precise localization of the various factors along the promoter DNA. Our results suggest key roles both for the mobile clamp of RNAP II and for TFIIE during the preinitiation stage of transcription.
|
|
|---|
and TFIIEß (37, 41, 52), and calf thymus RNAP II (20) were prepared as previously described. In-gel photo-cross-linking. The photoreactive nucleotides azidobenzoic (AB)-dUTP and AB-(Gly)2-dUTP were synthesized as described by Bartholomew and colleagues and Persinger and Bartholomew (1, 40). The in-gel photo-cross-linking procedure was performed as recently described (12). Briefly, various photoprobes placing one or two photonucleotides in juxtaposition to one (or a few) radiolabeled nucleotide at a specific position along promoter DNA were prepared enzymatically and gel purified. Each probe was incubated for 30 min at 37°C with highly purified RNAP II and recombinant general initiation factors (see below and figure legends for details) and then loaded on a polyacrylamide-N,N'-bisacryloylcystamine native gel (35). The gels were run in a cold room and irradiated with UV light (254 nm), and the complexes were localized using a PhosphorImager (Molecular Dynamics). Gel slices containing the complexes were excised and solubilized with dithiothreitol (0.4 M final concentration). The cross-linked complexes were then treated with S1 nuclease and DNase I prior to analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as previously described (44). Immunoprecipitation of the cross-linking products used specific antibodies directed against the various subunits of the general transcription factors as previously described (44).
Initiation assay.
Abortive initiation assays were done essentially as previously described (30, 39) with the following modifications. Double-stranded oligonucleotide templates (12 ng) carrying the adenovirus major late promoter from -45 to +35 were incubated for 60 min at 30°C with TBP (60 ng), TFIIB (30 ng), RAP74 (65 ng), RAP30 (30 ng), RNAP II (165 ng), TFIIEß (40 ng), and, when indicated, TFIIE
(60 ng) in 20 µl of a reaction mixture containing 750 µM ATP, 750 µM CTP, 10 µM UTP, 2.5 µCi of [
-32P]UTP, 12.5 mM MgCl2, 3 mM EGTA, and 0.84 U of RNase inhibitor/ml. After the reaction was stopped, the mixture was treated with calf intestine alkaline phosphatase (8 U) in order to reduce the background caused by free radiolabeled nucleotides. Transcripts were analyzed on a 23% polyacrylamide denaturing gel containing 7 M urea and quantified using a PhosphorImager.
Computer modeling. Prediction of low-energy AB-dUMP and AB-(Gly)2-dUMP conformers was performed using the Sybyl Molecular Modeling Package (version 6.4; Tripos Associates, St. Louis, Mo.). The models (see Fig. 6) were built by importing the RNAP II (accession code, 1I50) and TBP-TFIIB-DNA (accession code, 1Vol) PDB files from SWISS-PDB version 3.7 (17) and VMD version 1.7.2 (24) into the Cinema 4D XL Modeling Package (version 6.3; Maxon Computer, Thousand Oaks, Calif.).
![]() View larger version (66K): [in a new window] |
FIG. 6. Topological model of the RNAP II preinitiation complex. (A and B) Top (A) and side (B) of the proposed trajectory of the promoter DNA around the mobile clamp of RNAP II. The path of promoter DNA has been drawn to account for our in-gel photo-cross-linking results (see the text). (C and D) Locations of TFIIF (C) and TFIIE (D) subunits according to our photo-cross-linking results. For each subunit, we obtained cross-linking to two distinct promoter regions (dashed circles). Although our results suggest that two copies each of TFIIE and TFIIF subunits are present within the complex, we cannot rule out the possibility that these factors have two (or more) distinct domains held together by an unfolded linker.
|
|
|
|---|
, TFIIEß, and highly purified RNAP II were mixed with the various photoprobes, and the resulting complexes were separated on native polyacrylamide gels (Fig. 2A). The gels were irradiated with UV light, and gel slices containing the protein-DNA complexes were excised. Using this procedure, any intermediate complex that may have formed was disregarded. After dissolution of the gel fragments, the samples were treated with nucleases and the cross-linked polypeptides were then analyzed by SDS-PAGE. The cross-linked polypeptides were visualized by autoradiography of the gels and identified according to their molecular weights. The identities of the cross-linked polypeptides were confirmed either by immunoprecipitation of the cross-linking products with specific antibodies or by using tagged factors in the cross-linking reactions. The results of at least four independent experiments using each individual photoprobe were compiled to produce Fig. 3, which shows the cross-links that were reproducibly obtained. In each case, a control reaction lacking TBP was assembled to ensure promoter binding specificity (Fig. 2A shows an example). We also performed gel mobility shift experiments using subsets of the general initiation factors and RNAP II to ensure that our conditions for complex assembly allowed the formation of a bona fide TBP-TFIIB-TFIIF-RNAP II-TFIIE-promoter complex (Fig. 2B).
![]() View larger version (39K): [in a new window] |
FIG. 1. Photoprobes used for in-gel protein-DNA photo-cross-linking. (A) The 12 photoprobes used in our analysis are schematically shown. The position of the photonucleotide AB-dUMP or AB-(Gly)2-dUMP (U) along the adenovirus major late promoter is indicated (positions -15, -10, and -5) in each case and is used to name each photoprobe. Mutations (N to T) used in some of our photoprobes are indicated. The positions of the TATA element (underlined and boldface) and initiation site (arrow) are indicated. (B) Low-energy conformers of AB-dUMP and AB-(Gly)2-dUMP in the DNA major groove. Molecular simulation was used to predict the various possible conformers in each case. The photoreactive nitrene is shown in blue.
|
![]() View larger version (25K): [in a new window] |
FIG. 2. In-gel photo-cross-linking of the RNAP II preinitiation complex. (A) Schematic representation of the in-gel photo-cross-linking procedure. Preinitiation complexes were assembled by mixing each photoprobe with RNAP II and the general transcription factors. The TBP-IIB-IIF-RNAP II-IIE-promoter complex was electrophoresed in native polyacrylamide-N,N'-bisacryloylcystamine gels and irradiated with UV directly in the gels. Gel slices containing the cross-linked complexes were recovered, solubilized, and treated with nucleases. The resulting cross-linked products were then analyzed by SDS-PAGE. EMSA, electrophoretic mobility shift assay. (B) Gel mobility shift experiments showing the various intermediate complexes assembled using subsets of the general initiation factors and RNAP II. The positions of the various complexes are shown.
|
![]() View larger version (14K): [in a new window] |
FIG. 3. Summary of our in-gel photo-cross-linking results with a TBP-TFIIB-TFIIF-RNAP II-TFIIE-promoter complex. The boxes indicate the positions where each individual factor was found to cross-link to promoter DNA. +++, ++, and + represent the average relative intensities of the cross-linking signals obtained for each photoprobe using AB-dUMP. For example, the cross-linking of RPB2 at position -19 is more efficient than that of RAP30. The ± signs for the cross-linking of TFIIEß to photoprobes -10, -9/-8, -5, and +1 indicate the uncertainty concerning their specificities, as nonspecific complexes made of multiple molecules of TFIIEß were formed on these photoprobes in the absence of TBP (see the text for details). The use of AB-(Gly)2-dUTP in our photoprobes led to cross-linking results that are qualitatively similar to those using AB-dUTP. The positions of the TATA element (boldface) and initiation site (arrow) are indicated.
|
20 by 20 by 10 Å) in the major groove of the DNA helix (Fig. 1B). The simulations also indicated that only a slightly larger space is probed by the nitrene of AB-(Gly)2-dUMP than by that of AB-dUMP. This result may explain the similarity of the cross-linking data obtained with both photonucleotides. The flexibility of both AB-dUMP and AB-(Gly)2-dUMP indicates that photoprobes containing a single photonucleotide could theoretically react with multiple polypeptides that approach the space occupied by the reactive nitrene. This situation is exacerbated for those photoprobes containing two photonucleotides. Promoter contacts by the RNAP II subunits RPB1, RPB2, and RPB5. The in-gel site-specific protein-DNA photo-cross-linking results are summarized in Fig. 3, and some representative data are shown in Fig. 4. RPB1 cross-linked to seven positions from nt -31/-29 to +26, RPB2 cross-linked to nine positions from nt -40/-39 to +26, and RPB5 cross-linked to two positions from nt +7/+8 to +14 (Fig. 4B). Notably, RPB2 cross-linked to the promoter at six consecutive positions between -19 and +1, indicating that this promoter region closely approaches the second-largest RNAP II subunit. Between +7/+8 and +14, the promoter cross-linked to both RPB1 and RPB5, while further downstream, between +18/+20 and +26, cross-links were obtained with both RPB1 and RPB2. RPB1 cross-linked to the TATA box (-31/-29) and RPB2 upstream of TATA at -40/-39. The region between -10 and +1, where promoter melting occurs in the presence of TFIIH and hydrolyzable ATP, makes only seven cross-links with the TBP-TFIIB-TFIIF-RNAP II-TFIIE complex, four of which are with RPB2.
![]() ![]() ![]() View larger version (112K): [in a new window] |
FIG. 4. In-gel photo-cross-linking data. (A) Examples of the results obtained in our photo-cross-linking experiments with photoprobes prepared using AB-dUMP. The position of each cross-linked polypeptide is indicated. (B) Gel showing that the polypeptide that specifically cross-links at positions +7/+8 and +14 comigrates with the RPB5 subunit of RNAP II as identified using mass spectrometry. (C and D) RAP30 (C) and TFIIEß (D) cross-link to photoprobes -40/-39 and +18/+20. The identities of the cross-linking bands were confirmed by immunoprecipitation of cross-linked polypeptides with antibodies (Ab) raised against RAP30 and TFIIEß. +, present; -, absent.
|
60 bp apart, has been confirmed by immunoprecipitation of cross-linked RAP30 using an antibody directed against RAP30 (Fig. 4C). Our photo-cross-linking data for TFIIF subunits indicates that the location of RAP74 is centered downstream of +1 whereas that of RAP30 is centered on the TATA box and immediately downstream of it. Additional contacts by RAP74 upstream of TATA and by RAP30 downstream of +1 may be the result of tight bending and wrapping of the promoter DNA in the complex and/or of the presence of two RAP74-RAP30 dimers in the preinitiation complex (see Discussion).
Promoter contacts by TFIIE subunits.
The small subunit of TFIIE, TFIIEß, cross-linked specifically to four positions downstream of +1 between +7/+8 and +26 and to two positions upstream of TATA between -40/-39 and -31/-29 (see below). The cross-linking of TFIIEß to nt -40/-39 and +18/+20, two positions
60 bp apart, has been confirmed by immunoprecipitation of the cross-linked polypeptide using an antibody raised against TFIIEß (Fig. 4D). Notably, TFIIEß cross-linked to the same photoprobes as the RAP74 subunit of TFIIF, suggesting that these two subunits are closely positioned in the preinitiation complex. TFIIE
also cross-linked to the promoter downstream of +1 between +7/+8 and +26; the cross-link to +26 was weak. Notably, TFIIE
cross-linked to nucleotide -10. These results indicate that TFIIE
is the only general initiation factor that closely approaches the promoter in the region where DNA melting is to occur in the open complex.
We also obtained cross-linking of TFIIEß to photoprobes -10, -9/-8, -5, and +1. However, and for a reason that remains to be determined, all four photoprobes supported the efficient formation of a low-mobility complex in the absence of TBP (data not shown). When these nonspecific complexes, assembled in the absence of TBP, were UV irradiated, gel purified, and processed through our cross-linking procedure, we obtained strong signals corresponding to TFIIEß, as if these nonspecific complexes contained an aggregate made of multiple molecules of TFIIEß. Because we obtained strong nonspecific cross-linking of TFIIEß in the absence of TBP, we cannot be certain that the weak TFIIEß cross-linking signals obtained in the presence of TBP at positions -10, -9/-8, -5, and +1 are specific. For this reason, the cross-linking of TFIIEß between -10 and +1 is considered tentative (Fig. 3).
Role of TFIIE
in transcription initiation.
TFIIE has previously been shown to stimulate promoter melting in vitro (22, 23, 39). The cross-linking of TFIIE
to nucleotide -10 at the upstream end of the transcription bubble in the open complex supports this role. If the contact by TFIIE
at position -10 is important to promoter melting, we would predict that TFIIE
plays a key role during transcription initiation and affects transcription in vitro when using templates carrying short single-stranded (premelted) DNA regions. To test this hypothesis, we used an in vitro initiation assay that minimally requires the presence of TBP, TFIIB, TFIIF, and RNAP II (Fig. 5). A short double-stranded oligonucleotide containing the adenovirus major late promoter has been used as a template and was shown to support accurately initiated transcription in vitro (30). Compared to a reaction containing TBP, TFIIB, TFIIF, RNAP II, and TFIIEß, the number of initiation events was slightly increased upon addition of TFIIE
(Fig. 5A, left). Notably, the presence of a 2-nt single-stranded DNA region at -8/-7 resulted in strong stimulation of initiation by TFIIE
(Fig. 5A, right). Alternating this premelted region between -11 and -1 revealed that maximal stimulation occurred in the upstream half of the region that was to form the transcription bubble prior to initiation and where TFIIE
cross-links to the promoter (Fig. 5B). These results support the notion that TFIIE
, which closely approaches promoter DNA at position -10, is important in that it stabilizes promoter melting and stimulates the initiation of transcription.
![]() View larger version (30K): [in a new window] |
FIG. 5. TFIIE stimulates transcription initiation. (A) All double-stranded (0/0) and artificially premelted (-8/-7) DNA templates were used in an abortive initiation assay in either the presence (+) or the absence (-) of TFIIE . The positions of the abortive transcripts are indicated (3 to 10 nt). Stimulation levels obtained through the addition of TFIIE are indicated. The positions of the first G residue (boldface) and the initiation site (arrows) are indicated. (B) The position of the 2-bp premelted region along the promoter DNA influences the stimulation of transcription by TFIIE . Templates lacking (0/0) or carrying (-1/-2, -3/-4, -4/-5, -6/-7, -7/-8, -8/-9, and -10/-11) a 2-nt premelted region were used. Stimulation by TFIIE is indicated. The error bars indicate standard deviations.
|
|
|
|---|
in the absence of TFIIH (compare Fig. 3 with reference 8, Fig. 1). Because the TBP-TFIIB-TFIIF-RNAP II-TFIIE-promoter complex studied here was run through a gel prior to protein-DNA photo-cross-linking and because such a complex does not assemble in the absence of TBP (or when a photoprobe with a mutated TATA box is used; data not shown), we are confident that the nonspecific interactions are eliminated. Notably, the length of the promoter region that is contacted by RNAP II, TFIIF, and TFIIE also remains basically unchanged compared to our previous results. More specifically, RNAP II cross-links to the DNA from -40 to +26, a 66-bp region that corresponds to 225 Å of B-form DNA. This expanse is significantly larger than the largest span of RNAP II, which is
140 Å. Furthermore, as we observed in our previous experiments, RAP74, RAP30, and TFIIEß come into close contact with the promoter both upstream of TATA and downstream of +1. Together, these results indicate that promoter DNA is wrapped in the preinitiation complex (Fig. 6 shows a model), fully supporting our previous conclusions. In the model shown in Fig. 6, the trajectory of the promoter DNA has been drawn to best account for our in-gel cross-linking results. Because the TBP-TFIIB-TFIIF-RNAP II-TFIIE-promoter complex is likely to be an intermediate between RPc (closed complex) and RPo (open complex), we do not expect the promoter DNA to be fully melted in the -9-to-+2 region, with the template strand entering the active site of the enzyme. Open-complex formation, which is obtained in the presence of TFIIH and ATP, is expected to produce an important conformational change in the complex, bringing the template strand of the promoter deep into the active-site cleft. In our topological model, the downstream end of the promoter follows the channel between the two "jaws," where it cross-links to RPB1, RPB2, and RPB5. Between +1 and -20, the DNA closely approaches RPB2 before making a right turn in the region of the TATA element, where TBP and TFIIB also cross-link weakly, to maintain close proximity to RPB1. Tight bending redirects the promoter toward a domain of RPB2 in the -40 region. Only this path of the promoter DNA can easily explain the various contacts observed by the RNAP II subunits. As shown in Fig. 6B, the deduced path of the promoter DNA produces a left-handed loop against the mobile clamp of RNAP II, suggesting a role for the clamp structure in preinitiation mechanisms. Such a left-handed wrapping around the clamp is expected to relax the DNA helix, as opposed to a right-handed wrapping that would induce compaction of the DNA helix. Relaxation of promoter DNA may play a role in producing a partial single-stranded DNA region necessary for the action of the single-stranded helicase of TFIIH during open-complex formation.
Cross-linking of TFIIF and TFIIE both upstream of the TATA element and downstream of the initiation site indicates that both factors are associated with the mobile clamp of RNAP II. Tight wrapping of the promoter DNA around the polymerase clamp may be responsible for contacts by these two factors to promoter positions >60 bp away from each other (e.g., simultaneous promoter contacts at positions -40 and +26). Alternatively, or additionally, the putative presence of two molecules of each subunit of both TFIIF and TFIIE may help to explain the distal promoter contacts. Both TFIIF and TFIIE were shown to exist as
2ß2 heterotetramers in solution. In support of this idea that two dimers of TFIIF are associated within the preinitiation complex, we have recently purified an RNAP II-containing complex using TAP-tagged RAP30 that contains both a tagged and an endogenous RAP30 molecule (C. Jeronimo, M.-F. Langelier, M. Zeghouf, M. Cojocaru, D. Baali, D. Forget, S. Mnaimneh, A. P. Davierwala, J. Pootoolal, M. Chandy, V. Canadien, B. K. Beattie, D. P. Richards, J. L. Workman, J. R. Hughes, J. Greenblatt, and B. Coulombe, submitted for publication).
The cross-link of TFIIE
to nucleotide -10 is of interest. TFIIE
contains a DNA binding activity that has been anticipated as being important to promoter melting. The region carrying the DNA binding activity is homologous to region 2 of prokaryotic sigma factors (37). Notably, and in full support of our results, region 2 of
A was found to contact position -10 in the RNAP holoenzyme-promoter complex (33). Stimulation of transcription initiation by TFIIE
, a phenomenon that is enhanced when the promoter is artificially premelted in the upstream portion of the melted region of the open complex, suggests that the -10 contact by TFIIE
stabilizes promoter melting, thereby stimulating initiation. Taken together, our results provide novel insights into the fine molecular details of transcriptional initiation mechanisms.
This work is supported by a grant from the Canadian Institutes for Health Research. M.-F.L. and V.T. hold studentships from the Natural Sciences and Engineering Research Council of Canada and the Fonds de la recherche en santé du Québec. B.C. is a senior scholar from the Fonds de la Recherche en Santé du Québec.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»