Kruppel associated box zinc finger proteins (KRAB-ZFPs) are transcriptional repressors that make up the largest family of transcription factors in mammals (>500 members in mice an humans). KRAB-ZFPs are also rapidly evolving. Yet strikingly little is known about the physiological functions of the majority of family members. These proteins consist of an N-terminal KRAB domain that mediates recruitment of the co-repressor proteins KAP1/Trim28 and the histone H3K9 methyltransferase SETDB1 and a variable number of C-terminal zinc finger domains that mediate sequence specific DNA binding. We and others have hypothesized that KRAB-ZFPs function as a type of slowly adaptive immune system that transcriptionally silences mobile genetic elements called endogenous retroviruses or ERVs, the pro-viral remains of ancient viral infections. Several lines of evidence point to a potential function of the KRAB-ZFP family in binding and silencing of ERVs. First, the number of KRAB-ZFPs in mammals tightly correlates with the number of ERVs and closely related LTR transposons. Second, the KRAB-ZFP family member ZFP809, was isolated based on its ability to bind to the primer binding site for proline tRNA (PBS-Pro) of murine leukemia virus (MuLV). Third, deletion of the KRAB-ZFP co-repressor KAP1 leads to transcriptional activation of many families of ERVs in mouse ES cells. We are therefore testing whether KRAB-ZFPs are ERV silencers in the mouse. We have performed extensive studies to explore the function of the KRAB-ZFP family member ZFP809 using a combination of genome-wide localization analysis (via ChIP-seq) and functional analysis (mouse knockouts). To determine the genome-wide binding sites of ZFP809 we performed ChIP-Seq analysis in ES and embryonic carcinoma (EC) cells. We have found that ZFP809 binds specifically to ERVs related to MuLV and other genomic sequences harboring a Primer Binding Site for Proline tRNA (PBS-Pro), an 18nt consensus element. Importantly, the binding sites for ZFP809 overlap with the binding of KAP1 and SETDB1, and are enriched with the K3K9Me3 mark, consistent with a model of ZFP809-mediated recruitment of these factors. To explore whether ZFP809 is required for ERV silencing, we have generated homozygous mutant mice and primary cell lines (ES and PMEFs). We have found that mice homozygous for mutant alleles of ZFP809 are viable, but display highly elevated levels of VL30 PBS-Pro ERV elements confirming our hypothesis that ZFP809 is required for ERV silencing. To more fully characterize the gene expression defects in ZFP809 mutant mice, we performed RNA-seq analysis from primary cells (PMEFs) and tissues (spleen) from ZFP809 knockout mice. We find a strikingly specific phenotype in these cells: up-regulation of VL30 Pro-PBS elements and a handful of genes adjacent to these elements in somatic tissues. Importantly, we show that this reactivation phenotype is associated with an altered chromatin state, most notably a complete loss of KAP1 and reduced SETDB1 binding to these ERVs, reduced H3K9Me3, and increased acetylation, consistent with a switch from a repressed to an active state. Furthermore we also demonstrate the appearance of small but significant fraction of completely unmethylated CpG sites on VL30 elements exclusively in ZFP809 KO tissues, suggesting that ZFP809-dependent recruitment of KAP1 and H3K9Me3 is not completely responsible for initiating, but instead stabilizing this methylated state. Several lines of evidence support that the critical window of ZFP809 function occurs around the time of implantation. First, ZFP809 binds to ERVs in ES cells. Second, ZFP809 protein levels are considerably higher in ES cells than somatic cells. Third, phenotypes in Zfp809 KO PMEFs cannot be rescued by re-expression of ZFP809. Fourth, conditional deletion of Zfp809 in PMEFs does not phenocopy the ERV reactivation phenotype in PMEFs derived from Zfp809 KO mice. This data suggests that Zfp809 is required to initiate, but not fully maintain, the stable heterochromatic silencing of ERVs during embryo development. It has been suggested that KRAB-ZFPs perhaps evolve in response to endogenization of retroviral elements, as an evolutionary adaptation to reduce genome instability. We tested this hypothesis by searching muridoea for the presence of ZFP809 orthologues and Pro-PBS containing target ERVs. We demonstrate that the appearance of ZFP809 in muridoea predates the appearance of ERVs presently targeted by ZFP809, suggesting ZFP809 evolved its DNA binding domain against invading (exogenous) retroviruses. Nonetheless our data provides a compelling case for a KRAB-ZFP that functions rather specifically to transcriptionally silence retroviruses via the recruitment of heterochromatin machinery in a mammalian organism.
Showing the most recent 10 out of 20 publications
