Limits...
Molecular architecture of the yeast Mediator complex.

Robinson PJ, Trnka MJ, Pellarin R, Greenberg CH, Bushnell DA, Davis R, Burlingame AL, Sali A, Kornberg RD - Elife (2015)

Bottom Line: The model shows the locations and orientations of all Mediator subunits, as well as subunit interfaces and some secondary structural elements.Segments of 20-40 amino acid residues are placed with an average precision of 20 Å.The model reveals roles of individual subunits in the organization of the complex.

View Article: PubMed Central - PubMed

Affiliation: Department of Structural Biology, Stanford University School of Medicine, Stanford, United States.

ABSTRACT
The 21-subunit Mediator complex transduces regulatory information from enhancers to promoters, and performs an essential role in the initiation of transcription in all eukaryotes. Structural information on two-thirds of the complex has been limited to coarse subunit mapping onto 2-D images from electron micrographs. We have performed chemical cross-linking and mass spectrometry, and combined the results with information from X-ray crystallography, homology modeling, and cryo-electron microscopy by an integrative modeling approach to determine a 3-D model of the entire Mediator complex. The approach is validated by the use of X-ray crystal structures as internal controls and by consistency with previous results from electron microscopy and yeast two-hybrid screens. The model shows the locations and orientations of all Mediator subunits, as well as subunit interfaces and some secondary structural elements. Segments of 20-40 amino acid residues are placed with an average precision of 20 Å. The model reveals roles of individual subunits in the organization of the complex.

No MeSH data available.


Exhaustiveness of sampling and robustness of cross-link data.(A) Comparison of localization density maps calculated from the ensemble of solutions (500 best-scoring models) for the entire sample of models, the first half, the second half, and for jackknifing modeling runs (where 10% of cross-links were randomly removed). (B) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models, the first half, the second half, and for jackknifing runs. (C, D, E) Precision of Cluster 1 solutions (diagonal values) and average RMSD between Cluster 1 solutions (off-diagonal values) computed for the four ensembles, considering the whole Mediator complex, the Middle module, and the Tail module.DOI:http://dx.doi.org/10.7554/eLife.08719.011
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4631838&req=5

fig2s4: Exhaustiveness of sampling and robustness of cross-link data.(A) Comparison of localization density maps calculated from the ensemble of solutions (500 best-scoring models) for the entire sample of models, the first half, the second half, and for jackknifing modeling runs (where 10% of cross-links were randomly removed). (B) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models, the first half, the second half, and for jackknifing runs. (C, D, E) Precision of Cluster 1 solutions (diagonal values) and average RMSD between Cluster 1 solutions (off-diagonal values) computed for the four ensembles, considering the whole Mediator complex, the Middle module, and the Tail module.DOI:http://dx.doi.org/10.7554/eLife.08719.011

Mentions: In step 3, we conducted a large number of independent modeling runs. The positions and orientations of rigid bodies and flexible strings of beads were repeatedly perturbed in an effort to satisfy the scoring function consisting of the excluded volume, sequence connectivity, EM, and cross-linking restraints. A total of 165,523 Mediator model configurations were produced. The 500 best-scoring models (solutions) were grouped on the basis of RMSD into four clusters (C1-4, Figure 2—figure supplement 3), the minimum number required to fully represent the main structural differences between the best scoring models. To confirm that we had sampled conformational space sufficiently to reach model convergence, we compared two independent halves of the solutions to each other and to the entire set, showing they were all similar to one another (Figure 2—figure supplement 4).


Molecular architecture of the yeast Mediator complex.

Robinson PJ, Trnka MJ, Pellarin R, Greenberg CH, Bushnell DA, Davis R, Burlingame AL, Sali A, Kornberg RD - Elife (2015)

Exhaustiveness of sampling and robustness of cross-link data.(A) Comparison of localization density maps calculated from the ensemble of solutions (500 best-scoring models) for the entire sample of models, the first half, the second half, and for jackknifing modeling runs (where 10% of cross-links were randomly removed). (B) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models, the first half, the second half, and for jackknifing runs. (C, D, E) Precision of Cluster 1 solutions (diagonal values) and average RMSD between Cluster 1 solutions (off-diagonal values) computed for the four ensembles, considering the whole Mediator complex, the Middle module, and the Tail module.DOI:http://dx.doi.org/10.7554/eLife.08719.011
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4631838&req=5

fig2s4: Exhaustiveness of sampling and robustness of cross-link data.(A) Comparison of localization density maps calculated from the ensemble of solutions (500 best-scoring models) for the entire sample of models, the first half, the second half, and for jackknifing modeling runs (where 10% of cross-links were randomly removed). (B) Comparison of localization density maps calculated for Cluster 1 for the entire sample of models, the first half, the second half, and for jackknifing runs. (C, D, E) Precision of Cluster 1 solutions (diagonal values) and average RMSD between Cluster 1 solutions (off-diagonal values) computed for the four ensembles, considering the whole Mediator complex, the Middle module, and the Tail module.DOI:http://dx.doi.org/10.7554/eLife.08719.011
Mentions: In step 3, we conducted a large number of independent modeling runs. The positions and orientations of rigid bodies and flexible strings of beads were repeatedly perturbed in an effort to satisfy the scoring function consisting of the excluded volume, sequence connectivity, EM, and cross-linking restraints. A total of 165,523 Mediator model configurations were produced. The 500 best-scoring models (solutions) were grouped on the basis of RMSD into four clusters (C1-4, Figure 2—figure supplement 3), the minimum number required to fully represent the main structural differences between the best scoring models. To confirm that we had sampled conformational space sufficiently to reach model convergence, we compared two independent halves of the solutions to each other and to the entire set, showing they were all similar to one another (Figure 2—figure supplement 4).

Bottom Line: The model shows the locations and orientations of all Mediator subunits, as well as subunit interfaces and some secondary structural elements.Segments of 20-40 amino acid residues are placed with an average precision of 20 Å.The model reveals roles of individual subunits in the organization of the complex.

View Article: PubMed Central - PubMed

Affiliation: Department of Structural Biology, Stanford University School of Medicine, Stanford, United States.

ABSTRACT
The 21-subunit Mediator complex transduces regulatory information from enhancers to promoters, and performs an essential role in the initiation of transcription in all eukaryotes. Structural information on two-thirds of the complex has been limited to coarse subunit mapping onto 2-D images from electron micrographs. We have performed chemical cross-linking and mass spectrometry, and combined the results with information from X-ray crystallography, homology modeling, and cryo-electron microscopy by an integrative modeling approach to determine a 3-D model of the entire Mediator complex. The approach is validated by the use of X-ray crystal structures as internal controls and by consistency with previous results from electron microscopy and yeast two-hybrid screens. The model shows the locations and orientations of all Mediator subunits, as well as subunit interfaces and some secondary structural elements. Segments of 20-40 amino acid residues are placed with an average precision of 20 Å. The model reveals roles of individual subunits in the organization of the complex.

No MeSH data available.