Topographic map of the transcriptional factor NF-κB in the human genome

NF- κB is one of the most important transcription factors in mammals and participates in a plethora of biological processes, like inflammation, immunity, cell cycle regulation and apoptosis (Gilmore TD, 2006, Karin M, Ben-NeriahY, 2000). The question posed in the present thesis, was the study of the mechanism through which the transcription factor NF-κB can discriminate its targets, among a total of potential binding sites (Misteli T. 2001). In the beginning, by applying simple statistic methods, we were able to estimate that those potential binding sites in the human genome were about 1 million. This clue, combined with the fact that in most cells, NF-κB molecules are around 50.000, while the number of well known target genes of the factor, are approximately 500, leads us to wonder about what could be the functional role of those binding sites and how are they implicated in the regulatory action of the factor. In order to approach this initial question, we mapped the binding sites of 'F-κB in the human genome, by using unbiased methods. For this reason, we created a library with binding sites of the factor, after infection of the cells with Sendai virus for six hours. From all of the five proteins in NF-κB family, we chose to study the distribution of p65 subunit, because it participates in the most important heterodimer, the p50/p65('F-B) which is at the same time the most abundant and of inducible form in all cell types. This is the first study that shows positional distribution of the specific transcription factor through all the human genome, after inducing the cells with virus. In the past, a study about NF-κB (Martone R et al, 2003), had unraveled the binding sites of the factor only in the context of chromosome 22. Our results showed that 10% of the clones were known targets of NF- κB (e.g. troponin, NCAM1), while the rest 90% are obviously, potential new targets (unknown before this study) of the factor (e.g. TNFSF9, RelA). Based on the results of our analysis, we looked for the 10 basepair binding site of NF-κB that applies to the consensus GGGRNNYYCC. Through our search, we noticed the presence of another 10 basepair sequence that came up rather often but had a small variation from the original consensus. (~12% of library clones). 12 More specifically, this slightly different sequence is the GGGTTTCACC, which is 20% different, from the general consensus (underlined are the bases that make the difference). We also noticed that this sequence was exclusively present in Alu elements and in particular, only in AluSx, AluSg and AluY family members. Next, we studied the abundance of this specific sequence in the human and mouse genomes. The specific sequence, is present 170.000 times in the human genome, while if we change the central T residue to A, then the abundance frequency drops to the expected levels of any 10 base pair sequence, which is 3.000 times in the human genome (the conversion !1#, transforms the sequence, not to another NF-κB sequence, but to a random sequence). The same study in mice, showed no enrichment of this specific sequence, which implies that it is specific for primates. As far as the distribution of the clones in the genome is concerned, we noticed that a significant percentage of the clones are either in regions away from known genes, or in introns and only a small number is located in regions less than 10.000 bps upstream of known genes. A 12% of the library clones, turned out to have the element that we call Alu (GGGTTTCACC), which in turn, is present in more than 50% of clones that overlap with Alu elements (20% of all clones). We performed in vitro ChIP experiments and showed that randomly collected genomic regions from our library, are real targets of NF-κB. However, we noticed that after induction of the cells with TNF, the binding pattern of 'F-κB in the genome is altered. Also, this genomic regions may show random acetylation patterns of 03 and 04 histones, while in some of them we have found binding of TBP and RNA pol 44. Additional in vitro analysis with EMSA, proved that NF-κB binds to the Alu element of the library clones, although with less strength, compared to the binding affinity that this factor has for the PRDII element, from the enhancer of the interferon % gene. With in vitro transcription experiments, and by using specific templates and nuclear extracts from Hela cells, we showed that Alubinding site, has a limited ability to induce activation of transcription through binding of NF-κB. We showed that this is due to NF-κB’s inability to reinitiate 13 transcription from the same promoter. Also, in vivo experiments of transient transfection in mammalian cells ( 293! ) showed, that when this sequence is combined with one or multiple PRDII elements, gains functional properties of a typical enhancer and thus, when found under the control of this combination, increases the levels of transcription from the LUC reporter gene. The results of the above study, enrich our knowledge about the transcriptional regulation mediated by NF-κB, with two new evidences. First, we revealed a new type of binding site in the genome with which the specific factor interacts and it wasn’t monitored before (Alu). Secondly, due to our finding about Alu, we disclosed that a significant percentage of binding sites for NF-κB (20%), resides in Alu repetitive elements. This particular observation has multiple implications in an evolutionary as long as in a functional manner. Conclusively, from our study, it became evident that Alu elements, probably in cooperation with other adjacent cis-regulatory elements, have created an extra level of regulation of NF-κB activity, after the entrance of the latter in the nucleus. The formation of this regulatory level seems to be specific for primates, since the Alu element, has no ‘enhanced’ frequency in other groups of organisms, beyond the frequency that any random 10base pair sequence has in them. This particular sequence, seems to have coevolved with members of the Alu family of repetitive elements and it was favored by natural selection. This fact, established it as an in cis regulatory element with high frequency in the primate genomes. All the evidence we took under consideration in this study, show us that most likely, this binding sites serve as intermediate docking sites that deliver NF-κB at the cis-regulatory elements of the target gene