|Home | About | Journals | Submit | Contact Us | Français|
A shrewd cryo-EM study of the TFIID complex in this issue of Structure (Elmlund et al., 2009) has generated a markedly improved understanding of its structure and conformational dynamics. Accurate localization of TBP and other critical components and a new understanding of TFIID interaction with promoter DNA answer some significant questions and pose interesting new ones.
Transcription factor IID (TFIID), a large multi-protein assembly formed by the TATA box binding protein (TBP) and ~14 additional TBP-associated protein factors (TAFs), is a critical constituent of the transcription machinery responsible for the exquisitely regulated transcription of eukaryotic protein-encoding genes. TFIID plays an important role in recognition of promoter DNA and assembly of the preinitiation complex (PIC) and its biological significance has made it the target of extensive study (Cler et al., 2009). The size and complexity of TFIID have limited structural analysis to low resolution electron microscopy (EM) structures of the complex and high-resolution structures have been determined for only a handful of its individual components. Several EM structures of yeast and human TFIID complexes have revealed an overall horseshoe-shaped structure with a central channel that might include binding sites for TBP and DNA. Besides uncovering the overall topology of TFIID, previous EM studies had provided information about the likely location of TBP and some individual TAFs, and suggested that TFIID might exist in alternate conformations (Andel et al., 1999; Brand et al., 1999; Grob et al., 2006; Leurent et al., 2004). However, the precise structure of TFIID, the nature and origin of TFIID conformational variability and its possible functional significance, and the interaction of the complex with promoter DNA had until now not been characterized in detail.
Because single-particle EM analysis relies on combining information from thousands of individual particle images to overcome low signal-to-noise ratio limitations of individual 2D images and reconstruct a 3D structure, it is not surprising that the compositional and conformational heterogeneity of TFIID limited results from previous EM studies of the complex to rather low (30–40 Å) resolution. Through a combination of careful biochemistry and insightful EM image analysis Elmlund et al. (2009) now present an EM analysis of Saccaromyces pombe TFIID that greatly improves on previous work. The results from this comprehensive structural study advance considerably our understanding of TFIID structure and organization and provide tantalizing clues about correlations between the conformational dynamics and function of the complex.
Elmlund and his collaborators first pursued ab initio determination of a TFIID 3D structure from images of frozen-hydrated S pombe TFIID particles. This resulted in an initial TFIID reconstruction resembling previous TFIID structures in both overall shape and resolution. However, Elmlund et al. surmised that the resolution of their initial structure (~25Å) was too low considering the amount and quality of the data used to determine it. They suspected that heterogeneity of the complex might be a limiting factor in their analysis and implemented data analysis strategies that aimed to overcome this limitation and identify potentially different forms of TFIID. In brief, the authors assumed that the initial assignment of image orientation parameters used to calculate their initial TFIID reconstruction was generally correct but that images arising from different conformations of the complex were present for each orientation, thereby limiting the resolution of the reconstructed EM volume. They proceeded to use well established statistical methods to determine if images assigned to a given projection direction were heterogeneous and if they could be separated into homogenous subsets. This turned out to be the case, with images in each orientation-determined set clustering into two separate subsets. The authors then used a modified implementation of a previously published common-lines based protocol to assign subsets in each orientation to one of two overall sets of projections used to reconstruct two different TFIID structures which could be separately refined to better than 10Å resolution.
Although these two TFIID structures differed, large corresponding segments could be identified and one of the structures appeared to have an additional region corresponding in size and overall shape to TBP. Although TBP is generally regarded as the central component of TFIID, TBP-free forms of the complex have been characterized and it has been known that TBP is dynamically associated with yeast and human TFIID and that integrity of TFIID does not appear to be critically dependent on TBP binding (Sanders et al., 2002). After this, the story takes several interesting turns. The two TFIID structures in fact corresponded to TBP bound and TBP-free forms of TFIID, with particles in the EM samples nearly equally distributed between TBP-associated and TBP-free forms. Moreover, incubation of the purified TFIID with a large molar excess of TBP resulted in a majority of TFIID particles switching to the TBP-associated form.
The proposed rationalization of the two TFIID conformations that were observed is bolstered by further consideration of the TFIID EM volumes. The limited resolution of previous EM reconstructions of TFIID prevented docking of high-resolution structures available for some individual TFIID components. However, the much improved detail apparent in the reconstructions now reported by Elmlund et al. allowed the authors to reliably dock an NMR structure of TBP in complex with a portion of TAF1 into their larger TFIID volume. This allowed for precise localization of TBP to a somewhat peripheral position near the central TFIID pore that is consistent with TBP’s dynamic association with other TFIID components and with a previous antibody-labeling EM study that pinpointed TBP in the same general area. Interestingly, TBP appears to interact with other TFIID components in an orientation in which TAFs occlude the DNA-interacting surface of TBP, an observation that could have important implications for the mechanism of TFIID action.
The improved quality of the new TFIID structures also made possible the detection and interpretation of a large-scale TFIID conformation rearrangement that accompanies TBP dissociation. Fortuitously, this conformational change involves a remarkable rearrangement of two other TFIID components (TAF4 and TAF12) whose structures have been determined at high resolution. Docking of the X-ray structure of a human (TAF4-TAF12)2 complex into the TFIID reconstructions showed that dissociation of TBP leads to a rearrangement of the (TAF4-TAF12)2 complex.
Finally, the authors investigated the interaction of TFIID with promoter DNA, which seemed to associate preferentially (or at least more stably) with the TBP-free form of TFIID. Interaction of promoter DNA with TFIID in the absence of TBP induced a right-angle bending of the DNA and resulted in an extensive nucleic acid-TFIID interface. In agreement with results from previous biochemical analysis and with occlusion of the DNA-binding surface of TBP in the TBP-containing form of TFIID, the interaction of promoter DNA with TFIID seems to imply that disruption of TBP-TFIID contacts would be a prerequisite for interaction of promoter DNA with TBP. Also interesting is the observation that the geometry of the TFIID-promoter DNA interaction would cause TFIID to interfere with PIC assembly, in agreement with results from biochemical analysis of PIC assembly on immobilized DNA template assays (Ranish et al., 1999). Like other insightful reports, the analysis of TFIID presented by Elmlund et al. brings up perhaps as many questions as it addresses and will likely motivate further experiments that will expand our understanding of transcription initiation regulation.
Skillful biochemistry and perceptive image analysis, combined with propitious behavior of the TFIID complex and the availability of atomic-resolution structures of key TFIID components, made possible a study of TFIID structure and structural dynamics that is exceptionally revealing. Extension of the methodology used by Elmlund and collaborators to EM analysis of complexes showing increased conformational or compositional variability (or comparatively smaller structural changes) could prove unfeasible. However, methodologies that could be used to understand systems with a more complex compositional or conformational landscape have been proposed (Spahn and Penczek, 2009).
The work by Elmlund and coworkers provides another example of the unique and remarkable power of EM analysis to provide insight into how the structure and dynamics of a macromolecular complex relates to its function (Brignole et al., 2009). It also underlines the importance of making careful consideration of image heterogeneity an integral part of macromolecular EM studies. Increasing evidence suggests that conformational and compositional variability are likely ubiquitous in macromolecular assemblies and EM is uniquely positioned to reveal their significance. The development and implementation of novel and powerful image analysis methodologies will allow future practitioners of EM to generate vital information about the behavior of complexes responsible for a variety of cellular processes. Leveraging of EM structural information by consideration of atomic-resolution structures of individual complex components and correlation to biochemical, functional, and in vivo studies will undoubtedly prove essential for unraveling the mechanism of biomolecular machines.
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.