Skip to main content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Structure. Author manuscript; available in PMC 2010 Apr 15.
Published in final edited form as:
Structure. 2009 Apr 15; 17(4): 559–567.
doi: 10.1016/j.str.2009.01.016
PMCID: PMC2673807
NIHMSID: NIHMS100261
PMID: 19368889

Mediator Structural Conservation and Implications for the Regulation Mechanism

Gang Cai,1 Tsuyoshi Imasaki,2 Yuichiro Takagi,2 and Francisco J. Asturias1,*
Author information Copyright and License information PMC Disclaimer

Associated Data

Supplementary Materials

SUMMARY

Mediator, the multi-subunit complex that plays an essential role in the regulation of transcription initiation in all eukaryotes, was isolated using an affinity purification protocol that yields pure material suitable for structural analysis. Conformational sorting of yeast Mediator single particle images characterized the inherent flexibility of the complex and made possible calculation of a cryo-EM reconstruction. Comparison of free and RNA polymerase II (RNAPII)-associated yeast Mediator reconstructions demonstrates that intrinsic flexibility allows structural modules to reorganize and establish a complex network of contacts with RNAPII. We demonstrate that, despite very low sequence homology, the structures of human and yeast Mediators are surprisingly similar and the structural rearrangement that enables interaction of yeast Mediator with RNAPII parallels the structural rearrangement triggered by interaction of human Mediator with a nuclear receptor. This suggests that the topology and structural dynamics of Mediator constitute important elements of a conserved regulation mechanism.

INTRODUCTION

The multi-protein Mediator complex, conserved throughout eukaryotic organisms, is critical for in vivo assembly and stabilization of the preinitiation complex and conveys regulatory signals to the basal transcription machinery (Kornberg, 2005; Malik and Roeder, 2000; Naar et al., 2001). Nearly 20 years after Mediator was first identified in the yeast Saccharomyces cerevisiae (Flanagan et al., 1991; Kelleher et al., 1990) and despite its paramount importance in transcription, structural understanding of the complex is limited. A low resolution structure of yeast Mediator calculated from electron microscope (EM) images of Mediator particles preserved in stain (Dotson et al., 2000) showed few features and revealed only the overall morphology of the complex. Nonetheless, comparison with the structure of the Mediator-RNA polymerase II (RNAPII) holoenzyme (Davis et al., 2002) revealed large-scale changes in Mediator structure related to interaction of the complex with polymerase and brought up the question of how structural changes might contribute to the regulation mechanism (Chadick and Asturias, 2005).

EM studies of human Mediator also detected changes in its structure resulting from interaction with activators (Taatjes et al., 2002) and nuclear receptors (Taatjes et al., 2004), again suggesting that structural rearrangements might be important for Mediator function. However, the low resolution of the Mediator reconstructions and the absence of structural information about the metazoan Mediator-RNAPII holoenzyme have prevented the detailed comparison of the Mediator and holoenzyme structures necessary to attain a better understanding of critical issues such as the significance of conformational changes in Mediator triggered by interaction with RNAPII and other components of the basal transcription machinery and the extent of conservation of structural and mechanistic properties of Mediator.

Statistical analysis of single Mediator particle images preserved in stain can result in a detailed description of conformational changes in the complex and pave the way for cryo-EM analysis in which single particle images are instantaneously frozen from physiologically relevant conditions and analyzed in the absence of staining or any constraints related to crystallization. Unfortunately, such studies have been hindered by lack of a robust purification protocol capable of producing sufficient amounts of soluble, homogeneous, and transcriptionally-active Mediator. We report here on a new affinity purification method that readily and reproducibly yields functional yeast Mediator. This has made possible conformational analysis of Mediator particles preserved in stain and design of a strategy to pursue refinement of a cryo-EM reconstruction of Mediator that shows its structure in detail. Comparison with a previous structure of the RNA polymerase II (RNAPII)-associated form of yeast Mediator and with the structure of human Mediator results in a precise identification of Mediator structural modules, a description of the motions these modules undergo to accommodate interaction with RNAPII, and demonstrates that the structure and rearrangement of Mediator that enables interaction with polymerase are conserved from yeast to humans.

RESULTS

Implementation of a protocol for affinity purification of yeast Mediator

We had previously established a Mediator purification procedure comprising a series of conventional chromatography steps followed by Ni-NTA affinity purification (Takagi et al., 2005). This protocol worked consistently but required multiple steps and a relatively large amount of starting materials (~ 1 kg of yeast cells), which slowed structural studies. The new protocol is a modification of the Tandem affinity purification (TAP) method (Puig et al., 2001) that uses an alternate epitope-tagging vector in which the calmodulin-binding peptide sequence in the TAP tagging vector was replaced by an oligohistidine tag, enabling us to utilize a Ni2+-NTA column for a second affinity purification step. Preparation of a whole-cell extract free of DNA fragments (Takagi et al., 2005) and the use of ammonium sulfate precipitation to enrich the Mediator-containing fraction prior to affinity purification are also important aspects of the new protocol (Fig. S1).

Purified Mediator fractions were assayed by SDS-PAGE (Fig. 1A) and Mediator activity in basal and activated in vitro transcription assays was measured to establish that the purified complex was functional (Fig. 1B). Finally, the quality of the Mediator fractions was assessed by direct examination of the structural integrity of Mediator particles in stained and cryo-EM samples (Fig. 1C).

An external file that holds a picture, illustration, etc. Object name is nihms100261f1.jpg
Purification and characterization of Yeast Mediator

(A) Silver stain SDS-PAGE analysis of purified Yeast Mediator. (B) Reconstituted transcription assay of Mediator fractions performed with purified proteins at 24 °C as described (Materials and Methods), with ~0.2 pmol of each affinity-purified Mediator as indicated, in the absence (lanes 1, 2, and 4) or presence (lanes 3 and 5) of recombinant Gcn4 protein. Transcripts from templates containing Gcn4-binding sites (pGCN4) or Gal4-binding sites (pJJ470) and G-less cassettes (G-) are indicated on the right. 32P-labeled fX/Hinf III DNA molecular weight markers (M) are shown in the left with sizes. Transcripts (~360 bp an ~ 260 bp) were separated in a 6% denaturing PAGE and revealed by autoradiography (top). The ~360 bp bands were quantified by Fluorescent image analysis (bottom). (C) EM images of Mediator particles preserved in stain (left) and in amorphous ice (right). High-contrast images in stain were used for evaluation of Mediator conformational variability. (D) Individual Mediator particles preserved in amorphous ice are indicated by the arrowheads. The bar represents ~200nm.

Conformational analysis of Mediator particles preserved in stain

Images of stained Mediator particles were analyzed to detect possible changes in Mediator conformation with the idea of using that information to design a strategy for refinement of cryo-EM data. To this end, 7600K images of Mediator particles in the predominant orientation apparent in stained samples (Asturias et al., 1999; Dotson et al., 2000) were subjected to reference-free alignment and classification (Penczek et al., 1994). Assessment of the resulting class averages (Fig. S2) indicated variability in the conformation of the complex that would complicate cryo-EM studies. The major source of large-scale structural variability in Mediator particles was related to changes in the position of the previously identified Head module (Asturias et al., 1999) that could be either separated from the top portion of the Mediator structure (the standard conformation) or abutted to it (Fig. 2A). Further classification focused on specific segments of the structure was used to obtain mode detailed information about their movement. The Head domain pivots on a central connection to the top of the Mediator structure, covering a range of ~45° (Fig. S3, and Movie S1). A small, elongated feature at the very top of the structure seems to maintain the same overall conformation but is flexibly connected and highly mobile (Fig. S4 and Movie S2). Finally, the central portion of the Mediator structure, comprising the largest portion of its mass, is largely stable in conformation as evidenced by reproducible structural details across different class averages, but its top-right portion appears to undergo limited rotational rearrangement (Fig. S5 and Movie S3).

An external file that holds a picture, illustration, etc. Object name is nihms100261f2.jpg
Yeast Mediator conformational variability and Cryo-EM reconstruction

(A) A selection of class averages obtained by statistical classification of images of Mediator particles preserved in stain showing variability in the position of the Head domain (labeled H in the top middle average) and the top of the Tail domain (labeled T in the top middle average). The number of images included in each class average is indicated by the number under each average. (B) Initial models for alignment of cryo-images by projection matching were generated by filtering a previous stain reconstruction of Mediator to ~10 nm and computationally varying the position of the Head domain (as indicated by the red outline). (C) Cryo-EM reconstruction of yeast Mediator calculated from images of ~9,000 particles carefully selected based on their conformational homogeneity. Curved black arrows in the Front view indicate the approximate magnitude and direction of variability in the position of the Head and middle portions of the structure. Apparent multiplicity of the top portion of the Tail domain results from significant mobility. Red stars identify a portion of the density that seems to correspond best to the shape and most common position of the tip of the Tail domain.

Cryo-EM analysis of Mediator

The results from conformational analysis of Mediator particles preserved in stain suggested a strategy for cryo-EM analysis in which initial models with different Head module positions would be used to identify and align particles with the same overall conformation in a competitive projection matching refinement. A previous reconstruction of Mediator calculated by applying the Random Conical Tilt method (RCT) to images of Mediator particles preserved in stain (Dotson et al., 2000) was low-pass filtered to ~100 Å and computationally manipulated to obtain two initial models of Mediator with the Head in the normal and collapsed positions (Fig. 2B). Because definitive identification of Mediator particles in cryo-samples was difficult and to avoid any bias related to particle selection, an initial cryo dataset including ~32,000 particle images was assembled from all particle-like features in a set of 185 micrographs and screened by running repeated rounds of reference free alignment and classification. Only image classes with clear structural features were kept and combined to obtain an edited data set including ~16,500 particle images. Then, the two initial models with different Head positions were used to identify and align particles with the same overall conformation through a competitive projection matching refinement scheme. A subset of ~7,500 Mediator cryo-images showed higher cross-correlation values to the model of the standard Mediator conformation and resulted in a cryo-EM reconstruction of Mediator (Fig. 2C) with a resolution of ~28 Å (Fig. S6).

The cryo-EM reconstruction of Mediator clearly resembles the original RCT reconstruction from stained particles (Dotson et al., 2000) and a projection of the cryo-EM structure matches in detail to the 2D class average calculated from stained Mediator particles (Fig. S7), corroborating that the cryo-EM reconstruction corresponds to the same overall Mediator structure. However, although the nominal resolution of the cryo-EM Mediator reconstruction is not dramatically higher, the volume is considerably more intricate and provides a new level of understanding of Mediator structural organization. Particularly noteworthy is the structure of the top portion of the complex, which was largely featureless in the previous Mediator reconstruction but now appears complex, with several interconnected segments. The mobility of the small curved domain at the top of the structure revealed by analysis of stained Mediator particles must be even higher in cryo-specimens because the single domain apparent in 2D and 3D images from stained specimens is replicated in the cryo-EM reconstruction in the different conformations the domain must sample.

Mediator structural modules and RNAPII-induced rearrangement

The much greater detail apparent in the cryo-EM Mediator reconstruction prompted us to make a comparison with the structure of Mediator in a previously published reconstruction of the Mediator-RNAPII holoenzyme calculated from images of specimens preserved in stain (Davis et al., 2002). The RNAPII portion of that holoenzyme structure is compact and has well defined boundaries that made it easily distinguishable from Mediator density. The RNAPII density was computationally removed from the holoenzyme reconstruction to obtain a structure of Mediator in its RNAPII-associated conformation. Comparing the structures of Mediator in its free and RNAPII-associated forms resulted in a precise determination of the boundaries, connectivity, and relative motion of segments of the Mediator structure that shift as relatively rigid units (Fig. 3A).

An external file that holds a picture, illustration, etc. Object name is nihms100261f3.jpg
Structural modularity of yeast Mediator

(A) Boundaries of Head (red), Middle (green), Tail (orange), and Arm (purple) modules were defined based on independent mobility of these portions of the structure as semi-rigid modules identified by comparison with the conformation of Mediator in the Mediator-RNAPII holoenzyme structure (Davis et al., 2002). Some of the density (enclosed by black ellipse) assigned to the Head module forms an important connection between the Head and the rest of Mediator complex and most likely corresponds to Middle module subunits that are closely associated to the Head module and move along with it. The black arrows indicate module movement triggered by formation of the Mediator-RNAPII holoenzyme. The positions of the Middle and Tail modules change very significantly as they rotate and slide past one another (at an interface indicated by the dashed black lines), while the Arm module appears to rotate and move closer to the Middle and Tail modules. (B) Fitting of Mediator modules into the Mediator portion of a Mediator-RNAPII holoenzyme structure (semi-transparent gray surface). The Head module was fitted first and other modules were subsequently fitted as required to maximize overlap between the modules and Mediator density in the holoenzyme reconstruction. Further small rearrangement of specific features is suggested by the similarity between portions of the free and RNAPII-associated Mediator structures. For example, two small features (outlined in black) in the Tail module (third row from the top) match nearby features (outlined in red) in the Mediator portions of the holoenzyme structure. Apparent extra density at the top of the Tail module results from high mobility of that portion of the structure that results in apparent multiplicity of the feature. The Arm module seems to undergo significant conformational changes and could only be partially accommodated into the Mediator portions of the holoenzyme structure. Two fragments (outlined in black) of the Arm (bottom row) match nearby portions (outlined in red) of Mediator density in the holoenzyme structure but the Arm was not segmented to avoid over-fitting. (C) Mediator modules repositioned by comparison with the structure of the Mediator portion of the holoenzyme structure are shown along with the cryo-EM reconstruction of free Mediator (gray surface) to show the magnitude of the rearrangements undergone by Mediator modules upon interaction with RNAPII.

Four different Mediator structural modules were identified, computationally segmented from the cryo-EM reconstruction, and manually fitted into the structure of the RNAPII-associated (holoenzyme) form of Mediator. The Head module was fitted first, along with some additional density likely corresponding to Middle module subunits that seems to remain closely associated to the Head (Fig. 3B, first row). Other modules were fitted as required to maximize overlap between the modules and Mediator density in the holoenzyme reconstruction. In some cases, further small rearrangement of specific features is suggested by the similarity between portions of the two structures. For example, two small features (outlined in black) in the Tail module match nearby portions (outlined in red) of Mediator density in the holoenzyme structure (Fig. 3B, third row). Extra density at the top of the Tail module results from high mobility of that portion of the structure that results in apparent multiplicity of the feature. The Arm module seems to undergo significant conformational changes and could be only partially accommodated into the holoenzyme structure. Two fragments (outlined in black) of the Arm match nearby portions (outlined in red) of Mediator density in the holoenzyme structure but the Arm was not segmented to avoid over-fitting (Fig. 3B, fourth row). The magnitude of the changes in the relative position of Mediator modules triggered by interaction with RNAPII can be directly appreciated by comparing the cryo-EM reconstruction of free Mediator to the arrangement of the modules after docking into the RNAPII-associated Mediator structure (Fig. 3C).

Three of the new structural segments match the previously identified Head, Middle, and Tail modules. A fourth module (the Arm) was also identified thanks to the greater detail apparent in the cryo-EM Mediator reconstruction. The Head module is largely separate from the rest of the Mediator structure and its correspondence to the biochemically-defined Head module is evidenced by the similarity between its structure and that of an assembly of recombinant Head module subunits (Takagi et al., 2006). The connection between the Head module and the rest of the Mediator structure comprises density in the back of the Head (highlighted by circle in Fig. 3A) that is not apparent in the structure of the recombinant Head assembly (Takagi et al., 2006) and that must at least partially correspond to subunits that are not components of the biochemically-defined Head module. Changes in this portion of the Mediator structure appear to be critical for facilitating large-scale changes in Mediator structure.

A general correspondence between the structurally- and biochemically-defined Tail modules is demonstrated by analysis of a mutant Mediator lacking several Tail module subunits (Dotson et al., 2000) (Fig. S8) but the precise boundary between the Middle and Tail modules and the subunit composition of the newly-defined Arm module are uncertain, as these three modules are tightly associated in the Mediator structure. Although there seems to be overall agreement between the structurally- and biochemically-defined Mediator modules, structural modules might not exactly correspond to protein modules defined on the basis of biochemical and genetic interactions. In general, the close match between features in modules segmented from the cryo-EM reconstruction of Mediator and features in the Mediator portion of the holoenzyme structure (Fig. 3B) validates the structural details in these two completely independent Mediator reconstructions and makes possible a detailed analysis of the way in which the structure of Mediator is rearranged upon interaction with RNAPII.

Mediator-RNAPII interaction

The rearrangement of Mediator that results from interaction with RNAPII involves large changes in the relative position of Mediator modules. The Head, Middle, and Tail portions of the free Mediator reconstruction can be unequivocally docked into a reconstruction of the RNAPII-associated conformation of Mediator. No repositioning of the Arm module as a rigid unit can simultaneously optimize the match of all Arm subdomains. However, features in the Arm module can be matched between the free and RNAPII-associated forms of Mediator (Fig. 3B, bottom row), suggesting that the Arm is likely to undergo internal structural rearrangements likely facilitated by its extended organization (three globular segments joined by relatively thin connections). These rearrangements of Mediator modules seem to be facilitated by motions naturally present in the free form of Mediator. The pivoting motion of the Head module (Fig. S3, and Movie S1) and the rearrangement between the Middle and Tail modules (Fig. S5 and Movie S3) are both observed when Mediator interacts with RNAPII. The most prominent change in Mediator structure results from the relative rotation and translation of the Middle and Tail modules that leads to a complete repositioning of the Middle module (Fig. 3C). Once the modules are rearranged to match their positions in the RNAPII-associated form of Mediator the correspondence between features in the free Mediator structure and features in the RNAPII-associated form is very clear (Fig. 4A).

An external file that holds a picture, illustration, etc. Object name is nihms100261f4.jpg
Arrangement of Mediator modules in the Mediator-RNAPII holoenzyme

(A) Fitting of Mediator modules into the Mediator portion of a Mediator-RNAPII holoenzyme reconstruction (semi-transparent gray surface) shows the way the modules are positioned when Mediator interacts with RNAPII in an arrangement that in its general characteristics must be equivalent to the one that Mediator would adopt as part of a complete preinitiation complex. (B) Two views of a Mediator-RNAPII complex model based on consideration of the free- and RNAPII-associated forms of Mediator and a projection map of the Mediator-RNAPII complex (inset) that reveals contacts between the Head module and RNAPII subunits Rpb4/Rpb7. The interaction of RNAPII with Mediator is very extensive, as the latter surrounds the entire back face of polymerase (top) where the general transcription factors TBP and TFIIB presumably bind. Transcription factor TFIIF is also partially located in the back face of RNAPII and around polymerase subunits Rpb4/Rpb7, likely explaining a TFIIF requirement for stable interaction of Mediator (and the Head module in particular) with RNAPII. All of these contacts are consistent with reported stabilization of the preinitiation complex by Mediator. The interaction between the Mediator Head module and Rpb4/Rpb7 (bottom) could facilitate changes in Rpb4/Rpb7 position that could affect the conformation of the RNAPII clamp domain and therefore indirectly modulate interaction of the enzyme with promoter DNA.

The structure of the Mediator-RNAPII complex was analyzed in light of a much better understanding of the Mediator structure and a new projection map of the Mediator-RNAPII complex (Fig. 4B, inset) that shows the interaction of the two in more detail than was apparent in our previous study (Davis et al., 2002). The Head, Tail, and Arm Mediator modules make multiple and extensive contacts with much of the RNAPII surface along the back (upstream) face of the RNAPII structure and two adjacent sides (Fig. 4B, top), leaving free only the front (downstream) face of the enzyme. These multiple contacts were apparent in the original reconstruction of the Mediator-RNAPII complex (Davis et al., 2002) but their origin could not be conclusively determined. Comparison with the cryo-EM Mediator reconstruction now establishes that all of these contacts involve Mediator density. In fact, the reorganization undergone by Mediator modules results in a conformation that seems to maximize Mediator-RNAPII interaction. The Arm module in particular seems to undergo extensive rearrangement to accommodate interactions with RNAPII. Also of note is a close interaction between the Rpb4/Rpb7 polymerase subunit complex that appears to play a role in regulating the conformation of the RNAPII active site cleft (Armache et al., 2003; Armache et al., 2005; Bushnell and Kornberg, 2003) and the Head Mediator module (Fig. 4B, bottom).

DISCUSSION

Functional significance of Mediator pliability and RNAPII contacts

The cryo-EM reconstruction of yeast Mediator provides a much more detailed view of the complex than was afforded by a previous RCT reconstruction calculated from images of molecules preserved in stain (Dotson et al., 2000). Particle deformation in the stained samples provides a partial explanation for the limitations of the previous structure but conformational variability of Mediator, which was not taken into account in the previous analysis, likely played the major role in limiting the resolution of the Mediator RCT reconstruction. This assumption is supported by comparison of the new cryo-EM Mediator reconstruction with the Mediator portion of the previously determined RCT reconstruction of the Mediator-RNAPII holoenzyme (Davis et al., 2002). That comparison reveals complete correspondence between the structures of the free and RNAPII-associated forms of Mediator and implies that the many contacts between Mediator and RNAPII apparent in the holoenzyme structure must stabilize the conformation of Mediator.

Mediator subunits have an unusually high content of intrinsically disordered regions that most likely facilitate conformational pliability of the complex (Toth-Petroczy et al., 2008) and make possible the extensive conformational changes observed upon formation of the Mediator-RNAPII holoenzyme. The conformational changes undergone by Mediator have at least two results: they generate an RNAPII interaction site that is not present in the free conformation of Mediator, and they establish an extended network of contacts with RNAPII that involve a large portion of the RNAPII surface. Although the Head module can form a stable complex with RNAPII and TFIIF and might suffice to stabilize a minimal preinitiation complex (Takagi et al., 2006), the Head module alone is not sufficient to recapitulate Mediator function. This suggests that contacts between other Mediator modules and components of the basal transcription machinery are of critical functional significance.

Conserved structure and conformational changes in Mediator

Following the discovery of Mediator in yeast (Flanagan et al., 1991; Kelleher et al., 1990), it was suggested that similar complexes might not be present or be as important in higher eukaryotic cells where the mechanism of transcription regulation would likely be considerably more involved. However, Mediator complexes were soon identified in higher organisms based on limited but significant sequence homology of component subunits and their critical role in transcription regulation has been well established (Fondell et al., 1999; Gu et al., 1999; Ito et al., 1999; Rachez et al., 1999; Rachez et al., 1998; Ryu and Tjian, 1999; Sato et al., 2004). Class averages calculated from particles of yeast and human Mediators show considerable similarity, with details of internal Mediator structure matching between the two complexes (Fig. S9A). Incubation of Mediator with RNAPII under the relatively high (~400mM potassium acetate) salt concentration conditions used in this study seems to partially stabilize an intermediate in the process that results in formation of the Mediator-RNAPII holoenzyme. In this intermediate, RNAPII interacts in a CTD-dependent manner with a portion of the Mediator structure near the interface between the Middle and Tail Mediator domains. This is the same position reported for binding of the human RNAPII CTD to human Mediator (Fig. S9B), suggesting that the initial, CTD-dependent interaction of RNAPII with Mediator might proceed through a comparable series of steps in the yeast and human systems.

Finally, comparison of the cryo-EM reconstruction of yeast Mediator and the published structure of the thyroid hormone receptor-bound form of human Mediator (Taatjes et al., 2004) reveals considerable structural similarity between the two complexes. Corresponding domains can be identified and they seem to be comparably organized (Fig. 5). Most importantly, yeast and human Mediators seem to undergo similar conformational changes. Comparison of free and TR-bound human Mediator structures identified a major conformational rearrangement involving lateral movement of density that would correspond to the yeast Middle module (Fig. S9A and Fig. 5, bottom right) (Taatjes et al., 2004). That rearrangement in human Mediator mirrors the change in the position of the Middle module leading to generation of an RNAPII interaction site in the yeast Mediator. Interestingly, the binding site for the TR nuclear receptor corresponds in position to the hinge adjacent to the yeast Mediator Head module (Fig. 3A) that could be involved in facilitating structural rearrangement of the yeast Mediator. Given the limited sequence homology between yeast and human Mediators, these observations suggest that the structure and structural dynamics of Mediator are likely critical for a conserved mechanism of transcription regulation.

An external file that holds a picture, illustration, etc. Object name is nihms100261f5.jpg
Structural similarity between yeast and human Mediators

The structural similarity between yeast and human Mediators is evidenced by comparison of their 3D structures. Different views of the structure of thyroid hormone receptor (TR)-bound human Mediator (Taatjes et al., 2004) are remarkably similar to corresponding views of the yeast Mediator cryo-EM reconstructions. Moreover, the main change in human Mediator structure induced by TR binding (Taatjes et al., 2004) is the same as the main change required for interaction of yeast Mediator with RNAPII, namely, a swinging of the Middle domain from a central position (indicated by a black line) to a lateral position (indicated by a red line) towards the back of the Head module. This rearrangement opens a space that in the yeast Mediator-RNAPII holoenzyme accommodates RNAPII and allows the enzyme to interact with different Mediator modules. This suggests that structural conservation might be related to conservation of basic aspects of the mechanism of transcription regulation from yeast to human cells.

EXPERIMENTAL PROCEDURES

Construction of Affinity Tagged Yeast Strain

A novel epitope-tagging vector (10xHistidine-TEV-Protein A) was constructed by modifying the TAP tagging plasmid pBS1479 (Rigaut et al., 1999) to replace the calmodulin binding peptide (CBP) sequence with a 10xHis tag, yielding vector pYT6. The protease-deficient yeast strains BJ2168 (MATa leu2 trp1 ura3-52 prb1-1122 pep4-3 prc1-407 gal2) (Zubenko et al., 1980) was selected for Mediator tagging. A PCR-based genomic epitope tagging method (Schneider et al., 1995) was used to construct a yeast strain encoding a functional Mediator complex with different kinds of affinity epitopes. Firstly, three copies of the HA epitope were introduced into the C-terminus of the Med8 subunit (Takagi et al., 2005). Then, a modified (10xHistidine-TEV-Protein A) tag was fused to the C-terminus of the Med22 subunit, yielding strain CA001 (10xHis-TEV-Protein A-Med22, Med8-PreSci-3xHA).

Mediator Purification and in vitro transcription assays

Cells from the CA001 yeast strain were grown at 30°C to an O.D.600 of 4.0 in 2x YPD media. About 100 grams of cells were collected and a whole cell extract was prepared as previously described (Takagi et al., 2005). This whole-cell extract was selectively precipitated in 30–55% ammonium sulfate and re-suspended using 1 × TEZ buffer [50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 10 uM ZnCl2, 5 mM β-ME and protease inhibitors]. After the suspension was clarified by centrifugation the supernatant was incubated for 2 hours at 4°C with 1 mL of 50% slurry of IgG-sephrose resin beads (GE) that had been pre-equilibrated with 1×TEZ plus 250 mM ammonium sulfate. After incubation the beads were washed with 50 ml 1×TEZ plus 500 mM ammonium sulfate, followed by a second wash with 50 ml of 1×TEZ plus 50 mM ammonium sulfate. After equilibration of the column with 1×TEZ plus 100mM ammonium sulfate (without protease inhibitors), 100 units AcTEV protease (Invitrogen) were added to the resin beads and incubated overnight at 4°C. The Mediator complex was then eluted with 3 column volumes of 1×TEZ plus 100 mM ammonium sulfate, 10% glycerol was added and the resulting aliquot was snap-frozen in liquid nitrogen and temporarily stored at −80°C. For the next purification step, the IgG eluate fractions were thawed in ice and diluted 6–8 fold with Ni-NTA binding buffer [50 mM HEPES (pH 7.5), 0.01% NP-40, 10% Glycerol, 500 mM potassium acetate and 10 mM imidazole]. After clarification by ultracentrifugation, the supernatant was incubated with 200 μL of Ni-NTA resin beads (Qiagen) for 2 hr at 4°C, packed into a column, and drained by gravity. The resin beads were washed with 10 mL of washing buffer [50 mM HEPES (pH 7.5), 0.01% NP-40, 10% Glycerol, 500 mM potassium acetate and 40 mM imidazole] and the Mediator complex was eluted with 200 μL of elution buffer [50 mM HEPES (pH 7.5), 0.01% NP-40, 10% glycerol, 500 mM potassium acetate and 200 mM imidazole]. The final Mediator fractions were flash-frozen in liquid nitrogen after addition of 10% glycerol.

In vitro transcription assays using purified proteins were performed as described (Takagi and Kornberg, 2006). Quantification of transcripts on an absolute scale was performed using a FLA-5100 FUJIFILM fluorescent image analyzer and the Multi Gauge software package after addition of 1 nCi of α-32P UTP to the gel 5 min before the end of the run.

Electron Microscopy Sample Preparation and Data collection

Mediator aliquots (200 μg protein mL−1 in 400 mM potassium acetate, 20mM Hepes, 10% (v/v) glycerol, 1 mM DTT, 1 mM EDTA, 0.01% (v/v) NP40, and 1X protease inhibitors (pH of 7.5)) were diluted to a final concentration of ~30 μg protein mL−1 with a buffer containing 20 mM Tris, 500 mM potassium acetate, 1mM DTT, and 0.01% NP-40 (pH 7.5). About 3 μl of protein solution were applied to a carbon-coated Maxtaform, 300-mesh Cu/Rh EM specimen grid (Ted Pella, Inc., Redding, CA) freshly glow-discharged in the presence of amyl amine. Mediator particles were then preserved either by staining with a 2.0% (w/w) uranyl acetate solution or by flash freezing in amorphous ice (Dubochet et al., 1988).

Images from specimens preserved in stain were recorded under low-dose conditions using a Tecnai Spirit (Philips/FEI) microscope equipped with a LaB6 filament and operating at an accelerating voltage of 120 kV. Images were recorded on a Tietz (TVIPS GmbH) CCD camera at 42,000X magnification and approximately 1 μm underfocus. Cryo-EM images were recorded under low-dose conditions using a CM200 (Philips/FEI) microscope equipped with field emission gun and operating at an accelerating voltage of 120 kV. Images were recorded on Kodak SO-163 film, at a magnification of 50,000X and with underfocus values between 0.9 and 4.0 μm. Micrographs were digitized on a Zeiss/SCAI flat bed densitometer (ZI/Zeiss) using a step size of 7μm. Digitized images were 2-fold pixel-averaged, resulting in a final pixel size corresponding to 5.7 Å for the stained specimens and 4.6 Å for the cryo specimens.

Approximately 7600 images of stained Mediator particles were hand selected from 231 CCD frames and a total of 185 micrographs were digitized from the cryo-EM data set to yield ~32,000 Mediator images. Cryo-images were divided into groups accordingly to defocus values calculated independently for 12 distinct sections of every micrograph. All image analysis was carried out using the SPIDER software package (Frank et al., 1996).

Mediator structure flexibility analysis

Reference-free image alignment and classification were used to separate images of stained Mediator particles into groups homogeneous in conformation. Alignment parameters were further refined by applying a soft-edged mask covering the stable (central) portion of the Mediator structure to each image and repeating the reference-free alignment procedure. This resulted in considerably improved structural detail in the central, most stable portion of the Mediator structure, demonstrating the success of the alignment strategy. Focused classification of the optimally-aligned images was then used to characterize the behavior of highly mobile segments of the Mediator structure. A 2D mask covering a highly flexible region (Head module or top end of the Tail module) was applied to the aligned particle images and correspondence analysis (Borland and Vanheel, 1990) of the densities under the 2D mask followed by Hierarchical Ascendant Classification using the Ward’s criterion (Ward, 1963) were used to generate class averages with the Head and top of the Tail modules in different conformations. These class averages were then montaged into movies summarizing mobility of the domains in question.

Calculation of the Mediator cryo-EM reconstruction

A published reconstruction of Mediator (Dotson et al., 2000) calculated using the Random Conical Tilt (RCT) method (Radermacher, 1988) and images of Mediator particles preserved in stain was low-pass filtered to ~100 Å resolution and used to generate two starting model volumes to be used as references in a multi-reference projection matching refinement (Craighead et al., 2002; Gao et al., 2005). The two initial models differed by the position of the Head module (see Fig. 2B). Consistency of the reconstructed volumes with the cyro-EM data was monitored after each round of refinement by comparing re-projections of the volume to averages obtained directly from the original images by multiple rounds of reference-free alignment (Penczek et al., 1994) performed in image subsets defined by supervised classification.

Comparison of the free and RNAPII-associated Mediator structures

RNAPII density was segmented and erased from an RCT reconstruction of the Mediator-RNAPII holoenzyme using Chimera (Pettersen et al., 2004). Comparison of the Mediator cryo-EM reconstruction with the Mediator portion of the holoenzyme reconstruction resulted in identification of portions of the Mediator structure moving as rigid units. Each of these segments was then extracted from the cryo-EM Mediator reconstruction and manually fitted into the holoenzyme reconstruction using Chimera. Fitting of individual modules was optimized using the Fit Map Chimera command.

Supplementary Material

Acknowledgments

This work was supported by NIH grant R01 GM67167 to FJA and by an American Heart Association Scientist Development Award 0735395N and funding from the Sholwater Trust Fund to YT. TI is supported by a postdoctoral fellowship from the Human Frontier Science Program.

Footnotes

SUPPLEMENTAL DATA

Supplemental Data include nine figures and three movies and can be found with this article online at

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Armache KJ, Kettenberger H, Cramer P. Architecture of initiation-competent 12-subunit RNA polymerase II. Proceedings of the National Academy of Sciences of the United States of America. 2003;100:6964–6968. [PMC free article] [PubMed] [Google Scholar]
  • Armache KJ, Mitterweger S, Meinhart A, Cramer P. Structures of complete RNA polymerase II and its subcomplex, Rpb4/7. The Journal of biological chemistry. 2005;280:7131–7134. [PubMed] [Google Scholar]
  • Asturias FJ, Jiang YW, Myers LC, Gustafsson CM, Kornberg RD. Conserved structures of mediator and RNA polymerase II holoenzyme. Science. 1999;283:985–987. [PubMed] [Google Scholar]
  • Borland L, Vanheel M. Classification of Image Data in Conjugate Representation Spaces. Journal of the Optical Society of America a-Optics Image Science and Vision. 1990;7:601–610. [Google Scholar]
  • Bushnell DA, Kornberg RD. Complete, 12-subunit RNA polymerase II at 4.1-A resolution: implications for the initiation of transcription. Proceedings of the National Academy of Sciences of the United States of America. 2003;100:6969–6973. [PMC free article] [PubMed] [Google Scholar]
  • Chadick JZ, Asturias FJ. Structure of eukaryotic Mediator complexes. Trends in biochemical sciences. 2005;30:264–271. [PubMed] [Google Scholar]
  • Craighead J, Chang W, Asturias F. Structure of Yeast RNA Polymerase II in Solution. Implications for Enzyme Regulation and Interaction with Promoter DNA. Structure (Camb) 2002;10:1117–1125. [PubMed] [Google Scholar]
  • Davis JA, Takagi Y, Kornberg RD, Asturias FA. Structure of the yeast RNA polymerase II holoenzyme: Mediator conformation and polymerase interaction. Molecular cell. 2002;10:409–415. [PubMed] [Google Scholar]
  • Dotson MR, Yuan CX, Roeder RG, Myers LC, Gustafsson CM, Jiang YW, Li Y, Kornberg RD, Asturias FJ. Structural organization of yeast and mammalian mediator complexes. Proc Natl Acad Sci. 2000;97:14307–14310. [PMC free article] [PubMed] [Google Scholar]
  • Dubochet J, Adrian M, Chang JJ, Homo JC, Lepault J, McDowall AW, Schultz P. Cryo-electron microscopy of vitrified specimens. Q Rev Biophys. 1988;21:129–228. [PubMed] [Google Scholar]
  • Flanagan PM, Kelleher RJ, III, Sayre MH, Tschochner H, Kornberg RD. A mediator required for activation of RNA polymerase II transcription in vitro. Nature. 1991;350:436–438. [PubMed] [Google Scholar]
  • Fondell JD, Guermah M, Malik S, Roeder RG. Thyroid hormone receptor-associated proteins and general positive cofactors mediate thyroid hormone receptor function in the absence of the TATA box-binding protein-associated factors of TFIID. Proceedings of the National Academy of Sciences of the United States of America. 1999;96:1959–1964. [PMC free article] [PubMed] [Google Scholar]
  • Frank J, Radermacher M, Penczek P, Zhu J, Li Y, Ladjadj M, Leith A. SPIDER and WEB: Processing and visualization of images in 3D electron microscopy and related fields. J Struct Biol. 1996;116:190–199. [PubMed] [Google Scholar]
  • Gao N, Zavialov AV, Li W, Sengupta J, Valle M, Gursky RP, Ehrenberg M, Frank J. Mechanism for the disassembly of the posttermination complex inferred from cryo-EM studies. Molecular cell. 2005;18:663–674. [PubMed] [Google Scholar]
  • Gu W, Malik S, Ito M, Yuan CX, Fondell JD, Zhang X, Martinez E, Qin J, Roeder RG. A novel human SRB/MED-containing cofactor complex (SMCC) involved in transcription regulation. Mol Cell. 1999;3:97–108. [PubMed] [Google Scholar]
  • Ito M, Yuan CX, Malik S, Gu W, Fondell JD, Yamamura S, Fu ZY, Zhang X, Qin J, Roeder RG. Identity between TRAP and SMCC complexes indicates novel pathways for the function of nuclear receptors and diverse mammalian activators. Molecular cell. 1999;3:361–370. [PubMed] [Google Scholar]
  • Kelleher IRJ, Flanagan PM, Kornberg RD. A novel mediator between activator proteins and the RNA polymerase II transcription apparatus. Cell. 1990;61:1209–1215. [PubMed] [Google Scholar]
  • Kornberg RD. Mediator and the mechanism of transcriptional activation. Trends in biochemical sciences. 2005;30:235–239. [PubMed] [Google Scholar]
  • Malik S, Roeder RG. Transcriptional regulation through Mediator-like coactivators in yeast and metazoan cells. Trends in biochemical sciences. 2000;25:277–283. [PubMed] [Google Scholar]
  • Naar AM, Lemon BD, Tjian R. Transcriptional coactivator complexes. Annu Rev Biochem. 2001;70:475–501. [PubMed] [Google Scholar]
  • Penczek P, Grassucci RA, Frank J. The ribosome at improved resolution: new techniques for merging and orientation refinement in 3D cryo-electron microscopy of biological particles. Ultramicroscopy. 1994;53:251–270. [PubMed] [Google Scholar]
  • Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE. UCSF Chimera--a visualization system for exploratory research and analysis. Journal of computational chemistry. 2004;25:1605–1612. [PubMed] [Google Scholar]
  • Puig O, Caspary F, Rigaut G, Rutz B, Bouveret E, Bragado-Nilsson E, Wilm M, Seraphin B. The tandem affinity purification (TAP) method: a general procedure of protein complex purification. Methods. 2001;24:218–229. [PubMed] [Google Scholar]
  • Rachez C, Lemon BD, Suldan Z, Bromleigh V, Gamble M, Naar AM, Erdjument-Bromage H, Tempst P, Freedman LP. Ligand-dependent transcription activation by nuclear receptors requires the DRIP complex. Nature. 1999;398:824–828. [PubMed] [Google Scholar]
  • Rachez C, Suldan Z, Ward J, Chang CP, Burakov D, Erdjument-Bromage H, Tempst P, Freedman LP. A novel protein complex that interacts with the vitamin D3 receptor in a ligand-dependent manner and enhances VDR transactivation in a cell- free system. Genes & development. 1998;12:1787–1800. [PMC free article] [PubMed] [Google Scholar]
  • Radermacher M. The three-dimensional reconstruction of single particles from random and non-random tilt series. J Electron Microsc Tech. 1988;9:359–394. [PubMed] [Google Scholar]
  • Rigaut G, Shevchenko A, Rutz B, Wilm M, Mann M, Seraphin B. A generic protein purification method for protein complex characterization and proteome exploration. Nat Biotechnol. 1999;17:1030–1032. [PubMed] [Google Scholar]
  • Ryu S, Tjian R. Purification of transcription cofactor complex CRSP. Proceedings of the National Academy of Sciences of the United States of America. 1999;96:7137–7142. [PMC free article] [PubMed] [Google Scholar]
  • Sato S, Tomomori-Sato C, Parmely TJ, Florens L, Zybailov B, Swanson SK, Banks CA, Jin J, Cai Y, Washburn MP, et al. A set of consensus mammalian mediator subunits identified by multidimensional protein identification technology. Molecular cell. 2004;14:685–691. [PubMed] [Google Scholar]
  • Schneider BL, Seufert W, Steiner B, Yang QH, Futcher AB. Use of polymerase chain reaction epitope tagging for protein tagging in Saccharomyces cerevisiae. Yeast (Chichester, England) 1995;11:1265–1274. [PubMed] [Google Scholar]
  • Taatjes DJ, Naar AM, Andel F, 3rd, Nogales E, Tjian R. Structure, function, and activator-induced conformations of the CRSP coactivator. Science. 2002;295:1058–1062. [PubMed] [Google Scholar]
  • Taatjes DJ, Schneider-Poetsch T, Tjian R. Distinct conformational states of nuclear receptor-bound CRSP-Med complexes. Nature structural & molecular biology. 2004;11:664–671. [PubMed] [Google Scholar]
  • Takagi Y, Calero G, Komori H, Brown JA, Ehrensberger AH, Hudmon A, Asturias F, Kornberg RD. Head module control of mediator interactions. Molecular cell. 2006;23:355–364. [PubMed] [Google Scholar]
  • Takagi Y, Chadick JZ, Davis JA, Asturias FJ. Preponderance of free mediator in the yeast Saccharomyces cerevisiae. The Journal of biological chemistry. 2005;280:31200–31207. [PubMed] [Google Scholar]
  • Takagi Y, Kornberg RD. Mediator as a general transcription factor. The Journal of biological chemistry. 2006;281:80–89. [PubMed] [Google Scholar]
  • Toth-Petroczy A, Oldfield CJ, Simon I, Takagi Y, Dunker AK, Uversky VN, Fuxreiter M. Malleable machines in transcription regulation: the mediator complex. PLoS computational biology. 2008;4:e1000243. [PMC free article] [PubMed] [Google Scholar]
  • Ward JH. Hierarchical grouping to optimize an objective function. American Statistical Association Journal. 1963;58:236–244. [Google Scholar]
  • Zubenko GS, Mitchell AP, Jones EW. Mapping of the proteinase b structural gene PRB1, in Saccharomyces cerevisiae and identification of nonsense alleles within the locus. Genetics. 1980;96:137–146. [PMC free article] [PubMed] [Google Scholar]
-