What ENCODE is teaching us on the genomic sites of transcription factors

The binding to DNA of sequence-specific Transcription Factors -TFs- controls the deposition of histone marks and transcriptional activation. To do so, TFs tend to cooperate and synergize at multiple levels, including DNA-binding. Large scale projects, such as ENCODE, are investigating TFs genomic targets systematically. One of the emerging themes is that many TFs do not bind, or marginally bind in vivo to canonical sites previously well characterized by biochemical means in vitro. In addition, even for TFs with a canonical in vivo logo, a large cohort of their sites (20/40%) are devoid of it. ENCODE data indicate that NF-Y binding ovelaps significantly, often with precise stereoalignment, with numerous TFs involved in growth control, such as E2F, AP1 and E-box binding TFs. Some TFs lacking their canonical sites in their peaks in vivo show an abundance of NF-Y sites, with a visible theme of multiple CCAAT with precise spacing. This is quite relevant since (i) we and others have reported that p53 and p63 control certain growth promoting genes in the absence of canonical sequences and via tethering by NF-Y bound to multiple CCAAT boxes. (ii) Mutations in p63 and p53 alter the expression of such CCAAT-dependent promoters. We propose that full understanding of the strategy of TFs, including p53 family members and mutants of clinical relevance, will require the comprehension of binding to both their canonical and “unorthodox” locations in which other TFs are recruiting them onto DNA.