Basic Science Program, SAIC-Frederick, Center for Cancer Research Nanobiology Program NCI-Frederick, Frederick, MD 21702, U.S.A.Sackler Institute of Molecular Medicine, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel
Figure 1Simplified illustrations depict distinct stages of RNA Pol II transcription
(A) A possible nucleosome reordering process to make the unavailable promoter accessible. This is followed by PIC formation at the promoter region either via an activated recruitment or basal transcription accumulation. EC, elongation complex. (B) Atomic model of the PIC. The relative orientation between the TBP-bound DNA and Pol II is based on their interactions with TFIIB. The signature of transcription initiation is the formation of a transcription bubble with the DNA duplex melting in the Pol II active-site cleft. (C) Atomic model of the open PIC with the melted DNA drawn as broken lines. The open PIC complex then enters the elongation initiation stage to synthesize the first RNA nucleotide. The PEC continues to synthesize the nascent RNA and at a certain RNA length it dissociates from the promoter and leaves the scaffold for another quick re-initiation. At this stage, a mature EC completes the transcription, or a paused EC (not shown in the Figure) stalls at a proximal promoter region, waiting to be activated to resume transcription elongation. Pol II nomenclature is provided in [91,92].
Figure 2Schematic diagram illustrating two distinct mechanisms of transcription signal transduction from TF via Mediator to the transcription machinery
In mechanism I (A), the DNA-bound TF binds Mediator which in turn recruits GTFs/Pol II to complete transcription without interruption. In mechanism II (B), a stalled PEC resumes its processive elongation from a paused stage after the DNA-bound TF binds with Mediator. We assume that the RE is exposed and available for TF binding in both mechanisms, and that the core promoter may be available (nucleosome free) or unavailable for access in mechanism I but is always available in mechanism II. For simplicity and clarity only two GTFs (TFIIH and TBP), Pol II, Mediator, Spt5, NELF and TF are included in the drawing instead of all PEC members. Mechanism I denotes the concept that, following a binding event of the TF DNA-binding domain to the RE, the TF AD recruits Mediator through binding at the Tail/Middle modules. In contrast, in mechanism II PEC already occupies the promoter. After TF binds the RE, it binds to Mediator at the Head/Middle modules and the transcription signal which originates from the TF–RE interaction propagates through Mediator's Head to Pol II to disassociate NELF and enter processive elongation.
Figure 3Schematic two-dimensional organization map of the yeast Mediator and Pol II
The numbers relate to Mediator and Pol II subunits. The broken lines are disordered linkers. This organization is based on experimental data from crystal structures (subunits MED8–MED18–MED20 [36,37], MED7–MED31 and MED7–MED21 ); EM (Head module, [25,27]), the Middle and Tail modules have been described in previous studies [41,100]. The relative orientation between Mediator and Pol II has been described previously . Pol II Rpb4/Rpb7 and Clamp bind the Mediator Head module and the Pol II CTD binds the Middle module. In mechanism I, PEC formation is triggered by a TF AD-bound Mediator which interacts with Pol II (red) mostly via its Tail (pink) or Middle (blue) modules. In mechanism II, PEC is pre-assembled on the promoter and consequently there is a low level of Ser5 CTD phosphorylation. In both mechanisms, the recruitment of p-TEFb is needed to promote processive elongation after the RNA capping checkpoint. We speculate that the perturbation by AD binding to the Mediator Head module (green) causes a significant conformational change which propagates via MED17, MED18 and MED20 to Rpb4/Rpb7 and finally to the Clamp of Pol II. This propagation displaces NELF and facilitates the transition of a stalled PEC into an elongation phase. A large conformational change displayed by a sufficiently high population of Mediator will present ‘shape-shifting’. This map is for yeast; however, the human map is similar. The cartoon does not reflect the actual size of each of the Mediator subunits.
Figure 4Protein–protein interfaces of Mediator complexes whose X-ray structures are available
Mediator is a very large multi-protein complex composed of 10–25 subunits. High-resolution structural information on Mediator is extremely limited to small domain–domain interactions or small subunits, all of which constitutes only tiny fractions of a large complex. Protein complexes are represented as ribbons in two light colours with the interface highlighted in dark colours. Only interfacial side chains are shown in space-fill with hydrophobic residues coloured orange and hydrophilic residues in cyan. All three interfaces, MED18/20 (PDB code 2HZM), MED7C/21 (PDB code 1YKE), MED7N/31 (PDB code 3FBI), and MED11/22 (PDB code 3R84), clearly show the domination of hydrophobic interactions which reflect the characteristics of a two-state protein folding. In turn, this explains the specificity and the disorder nature of individual Mediator subunits. Folding-upon-binding transitions of disordered states lead to specific interactions and efficient signal propagation across interfaces, which is particularly important for large multi-chain complexes. Efficient signalling may explain why Mediator subunits are highly disordered. This Figure is available as an interactive three-dimensional structure at http://www.BiochemJ.org/bj/439/0015/bj4390015add.htm.
Table 1Mediator modules, subunits and interacting TFs and Pol II CTD
Mediator is a huge multi-protein complex (26 subunits in the human ). It is the primary regulator of the PEC, which includes TFIIA, TFIIB, TFIID, TFIIE, TFIIF, TFIIH, Mediator and Pol II [27,79]. Mediator is divided into Head, Middle and Tail modules. The interactions and functions of some subunits have been partly elucidated [24–27,79]. The Table lists some of the known TF AD interactions . Because only scant structural data are available, we have not distinguished between yeast and human. The CDK8 submodule triggers Mediator conformational change , MED5 has acetyltransferase activity, and Tail MED14 was shown recently to bind the N-terminal domain of PPARγ . The partial list in this Table illustrates that a TF can bind multiple Mediator subunits even belonging to different modules (e.g. p53, GCN4); however, a TF can also require different subunits, as in p53 activation of p21 by Nutlin which does not require CDK8, whereas UV light C does . Such cases could relate to communication pathways. Secondly, TFs do not necessarily require the same subunits. For example, while MED1 is not required for PPARγ , it is required by, for example, C/EBPβ , PGC-1α  and ERα . Yeast and human subunit interaction maps have been constructed [35,40,41]. The Pol II CTD has up to 52 heptapeptide repeats (YSPTSPS) . Ser5 and Ser7 are phosphorylated by CDK7 (TFIIH) which is recruited (with TFIIB) by Mediator–Pol II complex before transcription initiation . Ser2 phosphorylation by CDK9 is required for elongation. Ser5 phosphorylation is required for initiation of transcription elongation. Submodule CDK8 phosphorylates CDK7. Key mechanistic questions relate to how TF–Mediator binding activates Pol II and how this relates to the CTD phosphorylation requirement. C/EBPβ, CCAAT/enhancer-binding protein β; E1A, early region 1A; ER, oestrogen receptor; GABP, GA-binding protein; Gli3, GLI family zinc finger 3; GR, glucocorticoid receptor; HNF-4, hepatocyte nuclear factor 4; PGC-1α, PPARγ co-activator-1α; RTA, replication and transcription activator; Sox9, SRY (sex-determining region Y)-box 9; RAR, retinoic acid receptor; TR, thyroid hormone receptor.