Novel non-transposable-element regulation patterns of KZFP family reveal new drivers of its rapid evolution

Pan Shen; Qijian Zheng; Aishi Xu; Junhua Liu; Yushan Hou; Chao Gao; Chunyan Tian; Fuchu He; Dong Yang

potential target genes besides TEs. We found that old-zinc-finger-containing KZFPs The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.15.041848 doi: bioRxiv preprint Introduction 42 KRAB domain-containing zinc finger protein (KZFP) family, the largest family 43 of transcription factors in mammalian (Nowick et al., 2011) , has 363 member genes in 44 human genome. Generally, KZFP contains a KRAB domain and a C-terminal C2H2 45 zinc finger array with DNA-binding potential (supplementary fig. 1 ). Both C2H2 zinc 46 finger and KRI (KRAB Interior) motif, which is the ancestor of KRAB domain (Birtle 47 & Ponting, 2006) , are old motifs, appearing widely across animals, plants and fungi. 48 However, these two kinds of motifs did not appear in the same protein during the 49 lengthy process of evolution until their 'marriage' in the last common ancestor of 50 coelacanths and tetrapods (Imbeault, Helleboid, & Trono, 2017) about 400 million 51 years ago. After that, the KZFP family expanded and diverged quickly especially 52 during the evolution of mammalian. 53 The phenomena of the rapid evolution of KZFP family include three aspects: (1) 54 The strong specificity of KZFP family across species. Lots of species-specific KZFP 55 genes were reported in human, chimpanzee and other species (Huntley et al., 2006; The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.15.041848 doi: bioRxiv preprint 9 these KZFPs are completely structured ( fig. 2C ), suggesting that they have a single 166 and unchangeable function execution pattern and constantly interact with a specific 167 co-factor (such as KAP-1). In another evolutionary path, the key amino acids in the 168 zinc fingers are conserved (that is, KZFPs containing old zinc fingers). However, the 169 KRAB domains of these KZFPs are variable and tend to be disordered ( fig 2C) , 170 suggesting that their KRAB domains are versatile, interacting with multiple proteins. ; supplementary table 4) . 214 Furthermore, we confirmed that MACS2 used in this study is more accurate than