Eukaryotic transcription requires the concerted action of numerous proteins including activators, co-activators, the general transcription factors (GTFs), RNA polymerase II (RNA Pol II), chromatin remodelers and modifiers, and chromatin- binding effector proteins. Although many of the components of the preinitiation complex (PIC) have been described, a comprehensive understanding of all preinitiation components and what affects chromatin modifications have on PIC formation remains vague. This doctoral thesis investigates the composition and function of eukaryotic PICs in HeLa and mouse embryonic stem cells. Using Multi Dimensional Protein Identification Technology (MuDPIT), others in the lab showed that most transcription components necessary for initiation were present in in vitro formed PICs in both cell types. Additionally, numerous chromatin remodeling proteins, elongation factors, and chromatin modifying enzymes were also identified proteomically. Intriguingly, two major co-activator complexes SAGA and Mediator, ranked among the highest activator-inducible proteins as detected by MuDPIT analysis. I generated HeLa nuclear extracts immuno-depleted of Mediator and SAGA and performed mechanism-based immobilized template experiments on chromatin to show both complexes interact with GAL4-VP16 directly, and do not compete for promoter access. I also used the immobilized template to show that Mediator, but not SAGA, serves as the main scaffold to recruit the PIC machinery. Genome-wide, enrichment patterns of Mediator, RNA Pol II, TATA-binding protein, and mRNA levels strongly correlate suggesting a prominent role for Mediator in PIC assembly. Lastly, I used in vitro transcription assays to show that SAGA is required for activated transcription on chromatin. Overall, these data support a working model where Mediator serves as the main co-activator complex, contributing to PIC formation, stability, and function, while SAGA acts after PIC formation, conferring activity to PICs on chromatin.
Although the role of histone tail modifications and subsequent effector binding proteins is an area of considerable study, the role that histone modifications play in PIC formation and activity is poorly understood. This thesis examines the role of the Chromodomain Helicase DNA-binding protein 1 (CHD1), known to bind histone H3 trimethylated at Lysine 4, in PIC assembly and function. I purified murine CHD1 to near homogeneity using a baculovirus-based insect cell system, and prepared the samples used for identification of CHD1-interacting proteins by MuDPIT of associated proteins from HeLa nuclear extracts. My collaborator on this project, Justin Lin, used purified CHD1 to show that the protein stimulates transcription on chromatin and further activates transcription on H3K4me3 chromatin. Further, mechanistic experiments performed by others in the lab revealed that CHD1 recruitment to immobilized templates is dependent on Mediator and enhanced by H3K4me3. Our in vitro and in vivo experiments suggested that Mediator and CHD1 interact to enhance transcription. The summation of these data support a model where Mediator acts to support both PIC assembly and the binding of CHD1 to PICs, with H3K4me3 providing a second interaction platform responsible for the recruitment of CHD1 to active genes.
In an opposing role, chromatin modifications also play a role in establishing and maintaining a silent transcriptional state. In the final chapter of this thesis, I investigate the affects of Polycomb Repressive Complex 1 (PRC1) on PIC formation and activity. I purified PRC1 and investigated the role of the complex in PIC assembly and function on H3K27me3 templates. I show that PRC1 can both block and dissociate the majority of PIC components including Mediator, SAGA, elongation factors and the general transcription factors to silence transcription. Notably, the ability of GAL4-VP16 and TATA-binding protein (TBP) to bind promoters was retained in the presence of PRC1. I again used immobilized template experiments with purified Mediator, TFIID, and PRC1 to show that PRC1 specifically blocks Mediator but not TFIID. Utilizing previously reported genome-wide binding in mouse embryonic stem cells, we showed that Ring1b and TBP displayed enriched binding profiles at developmental genes, and this enrichment significantly correlated with reduced expression as compared to genes enriched for Mediator, TBP, and Ring1b. In summary, the results of this study highlight a possible role for TFIID and PRC1 in co-regulating the expression of developmental genes.