We have been studying regulatory elements that are single-stranded when the human c-myc proto-oncogene is expressed and have been characterizing the conformation and topology sensitive DNA binding proteins that interact with these elements. First, a cell type, differentiation specific positive cis-element, FUSE, 1.5 kb upstream of promoter P1 is single stranded when c-myc is expressed in vivo and is devoid of nucleosomes except when c-myc is repressed. This element specifically binds FBP, a sequence specific, single strand DNA binding protein. FBP's amino terminus is a potent transcriptional repressor which interacts with TFIIH while the carboxyl is a powerful transcription activator. The activation domain of FBP binds directly with the p62 subunit and the p89 helicase subunit of TFIIH. A co-repressor with FBP, FIR, binds the central domain of FBP and interacts with TFIIH. FIR increases the affinity of FBP for binding with the FUSE element. FBP activation is defective in xeroderma pigmentosum cells mutated in TFIIH subunits. So endogenous c-myc expression is unresponsive to FBP in XP cells. In XPB cells we have discovered that tight regulation of c-myc is lost and remarkable cell-to-cell variation in MYC levels occur. The XPB mutation prevents FUSE from looping and interacting with the major P2 promoter, as a result both the induction and shutoff of c-myc in response to serum addition and removal are impaired. Thus, the same mutation that disables nucleotide excision repair also deregulates a dangerous oncogene. Moreover we have discovered that this mutation also disturbs normal cell cycle progression and leads to an accummulation of cells in G2/M. Knockdown of FIR with siRNA provokes a similar cell cycle alteration and similarly disturbs the serum response of the c-myc gene. Along with FBP, two other family members FBP2 and FBP3 probably form a basis set to adjust the steady levels of important genes. The central domain of FBP mediates interaction with FUSE. This portion of the protein is composed of repeated """"""""KH"""""""" motifs which comprise a bi-partite DNA binding domain. Sub-domains constituted of the amino-terminal two KH repeats or the carboxyl pair of KH motifs bind weakly and strongly, respectively with upstream and downstream contiguous sequence segments of FUSE. Recent SELEX studies have identified specific DNA sequences interacting with FBP's KH domains and have enabled a genomic analysis of prospective FBP binding sites. Upon binding with FIR, the weak binding sub-domain of FBP recognizes the upstream segment of FUSE with increased affinity. Both FBP and FIR have been shown to control levels of endogenous c-myc expression. FIR repression of c-myc also fails in XP cells. ChIP (chromatin immunoprecipitation)-on-chip (microarray) studies are revealing that the FBP-FIR system is a commonly used molecular machine to control the output of a very large number of genes. NMR studies of a complex between the strong binding sub-domain and the downstream segment of FUSE suggest that FBP is truly a DNA binding protein as particular features of the complex are unique to DNA-protein interactions. X-ray crystallography in collaboration with Dr. Demetrios Braddock is revealing new features of the FBP-FIR-FUSE system. A third protein binds with activating segments in the carboxyl terminus of FBP. This protein also blunts FBP activation. All of these interactions occur both in vivo and in vitro. The mechanism of transcriptional modulation by the FBP/FIR/TFIIH/repressor/co-activator complex reveals that FBP hastens RNA polymerase movement through earliest stages of transcript elongation. In contrast FIR delays the transition of RNA polymerase into its elongation mode. mechanisms. The relevance of FBP and FIR to the regulation of c-myc has been established using RNAi. Additional experiments have exposed the presence of multiple signals embedded within FBP targeting the protein to multiple nuclear compartments. We are currently commencing analysis of conditional knockout mice of FBP and FIR. ChIP-chip experiments are in progress to map the binding sites of these factors across the genome. We have also developed a a c-myc-EGFP knock-in that is the best reporter system yet developed to sutdy the regulation and deregulation of c-myc expression. combined with our studies of FIR and FBP via DNA topology these studies provide an extra dimension to our understanding of gene regulation. Second, a complex set of factors binds to a cytidine-rich nonanucleotide sequence, repeated five times, found 100-150 bp upstream of P1. This region, required for transcription from P1 as well as augmenting expression from P2, binds several sequence specific, single strand or supercoil DNA binding proteins. hnRNP K binds to this CT-element in supercoiled DNA. In vitro transcription experiments and transfection studies indicate that hnRNP K participates in the stimulation of gene expression through the CT-element. Structural studies are underway to determine how hnRNP K binds with the CT-element. The properties of FBP and hnRNP K prompted examination of c-myc regulatory cis-elements in vivo and in vitro to determine how these sites become melted. Using actively transcribing T7 RNA polymerase to generate torque, recent experiments indicate that the dynamic transmission of mechanical stress destabilizes particular elements, such as FUSE and CT-elements, even in the absence of defined topological boundaries. Recent experiments have been devised to measure the level of superhelical stress transmitted into DNA by ongoing transcription. The level attained is close to the theoretical limit and is suprisingly high, high enough to disturb chromatin and DNA structures and so may of regulatory consequence. Recent studies are establishing the the equivalence of our in vitro and in vivo studies. Because single-stranded DNA is much more flexible to torsion and flexion than is duplex, interposing CT-elements between genetically interacting sites facilitates these interactions. Therefore one function of regulated single-stranded cis-elements is to serve as protein-DNA hinge. Recent work suggests that competition between hnRNP K and SP1 at the CT element may be mediated by a conformation switch
Showing the most recent 10 out of 11 publications
