1. KRAB-ZNF genes are transcriptional repressors that make up the largest family of transcription factors in mammals which are rapidly evolving. Yet strikingly little is known about the physiological functions of the majority of its members. These proteins consist of an N-terminal KRAB domain that mediates recruitment of the co-repressor protein KAP1/Trim28 and a variable number of C-terminal zinc finger domains that mediate sequence specific DNA binding. Several lines of evidence point to a potential function of the KRAB-ZNF family in binding and silencing of ERVs. First, the number of KRAB-ZNF genes in mammals tightly correlates with the number of ERVs and LTR transposons. Second, the KRAB-ZNF 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-ZNF co-repressor KAP1 leads to transcriptional activation of many families of ERVs in mouse ES cells. We will therefore test if KRAB-ZNF genes serve as ERV silencers in the mouse, focusing on two candidates. We have initiated preliminary studies to explore the function of two candidate KRAB-ZNF genes, Zfp809 and Zfp568. These candidates were identified based on studies linking these factors to retroviruses (Zfp809) or based on phenotypic similarities to KAP1 mutants (Zfp568). To determine the genomewide binding sites of Zfp809 and Zfp568, we performed ChIP-Seq analysis on ES and embryonic carcinoma (EC) cells expressing Flag-epitope tagged factors (Zfp809) or with polyclonal antibodies raised against endogenous proteins (Zfp809 and Zfp568). Based on preliminary computational analysis, we have found that Zfp809 binds specifically to ERVs related to murine leukemia viruses and other genomic sequences harboring a PBS-Pro consensus element (Figure 1A-C) whereas Zfp568 preferentially associates with a distinct family of ERVs called intracisternal A particles (IAPs) within their 5UTR. We will next validate these binding sites using ChIP-QPCR and EMSA approaches, and confirm that the binding profiles disappear when ChIPs are performed in knockout ES cells (characterized next.) To explore whether Zfp809 and Zfp568 are required for ERV silencing, we have generated/are generating homozygous mutant mice and ES cells. We have found that mice homozygous for a Zfp809 genetrap allele are viable, but display highly elevated levels of VL30 PBS-Pro ERV element expression and Gm9705 expression (both of which are bound by Zfp809, Fig 2A-B) in somatic tissues confirming our hypothesis that Zfp809 is required for ERV silencing (Figure 2D-E). To more fully characterize the gene expression defects in Zfp809 mutant mice, we will perform RNA-sequencing on primary cells and tissues from Zfp809 knockout mice. Furthermore, we will perform epigenome profiling of wild type and mutant somatic cells (DNA methylation and ChIP-seq) to look at histone H3K9Me3, a known repressive mark recruited by KAP1 complexes to characterize the potential mechanism of gene de-reperession. Since we have observed that the Zfp809 mutant mice still expresses low levels of Zfp809 (ie, it is a hypomorphic allele), we are also developing conditional Zfp809 null mice sing a floxed allele. This will also allow us to use tissue specific knockout to distinguish whether Zfp809 is required for the initiation of and/or for the maintenance of ERV silencing. Zfp568 conditional knockout mice are also being generated so we can profile the expression of ERVs in embryos and somatic tissues as well as perform epigenome profiling on ES cells derived from these mutants. We anticipate that each KRAB-ZNF will bind to a distinct set of ERVs, and that a subset of ERVs bound by each KRAB-ZNF will also be improperly activated in mutants, and that the reactivation will be caused by failure to properly establish and/or maintain repressive epigenetic marks. We also expect to find that some non-viral genes will fall under the regulation of each KRAB-ZNF, an idea that is supported by our identification of the Gm9075 gene, which is bound by Zfp809 via a divergent PBS-Pro site, and whose expression is highly activated in Zfp809 mutants (Fig. 2B, E). In sum, these studies would be the first to directly link KRAB-ZNF genes to the silencing of ERVs, and be among the first to describe both the genome-wide binding profiles and biological functions for members of this rapidly evolving gene family implicated in speciation, development, and disease. 2. One major question being explored by developmental biologists is how transcription factors regulate developmental gene expression, and how the chromatin environment can facilitate or impede transcription factor accessibility and activity. The finding that somatic cell lineages can be converted to iPS cells or other somatic cell types using cocktails of transcription factors demonstrate the power of transcription factor mediated reprogramming, but because the process is slow and rare suggests that chromatin may be a major barrier. Using mouse development and genetics as a model system, a number of groups have identified important transcription factors and chromatin regulators that are required for specification of cell types of interest. But in many cases it is difficult to determine the target genes of these factors genome-wide, since it is not easy to obtain pure populations of cells in large enough quantities. The cell type of interest to my group is the motor neuron (MN), since we know a number of the transcription factors that are important for its specification, and have reporters that allow us to isolate these cells. Furthermore, the motor neuron is the cell type most severely affected in ALS and SMA patients, giving it important clinical relevance. We have thus initiated studies that aim to address a number of fundamental questions regarding the regulation of developmental gene expression by transcription factors in MNs that overcome the limitation of the small quantity of cells in the developing mouse embryo. Following is a step-by-step introduction to the approach underway and what we have accomplished thus far, followed by a description of what comes next. We have generated ES cell lines where we can integrate transcription factors (one at a time or in combination) into a single locus (the HPRT locus) under the control of a dox inducible promoter with high efficiency. We have integrated three epitope tagged transcription factors as a single cistron into this locus (the bHLH factor Ngn2 N the lim homeodomain factors Isl-1 I and Lhx3 L, which have been shown to cause ectopic MN formation in the chick neural tube when overexpressed). Upon dox induction of the ES cell lines, the NIL factors are expressed, and motor neuron genes are activated within 12 hours, indicating that ES cells are a permissive cell type that can rapidly be converted to MN like cells with 3 txn factors. We have cloned all combinations of NIL (single factors, double factors, and all 3) and are integrating into the HPRT locus We will perform RNA-Seq and ChiP-seq after induction to look for target genes of these factors in single or in various combinations We have generated a mouse line with the inducible NIL cassette. We will test dox induction during mouse development to determine what cell types in the developing embryo are permissive to transcription factor mediated reprogramming and which are restrictive. We are generating PMEF lines from NIL inducible mice. These lines should represent a restricted cell type, based on the likely slow and less robust differentiation into MNs. We will test by ChIP-Seq whether the transcription factors have similar access to their target genes when compared with permissive ES cells.
|Belfort, Marlene; Marraffini, Luciano; Macfarlan, Todd et al. (2014) Mobile DNA: an evolving field. Mob DNA 5:16|
|Gifford, Wesley D; Pfaff, Samuel L; Macfarlan, Todd S (2013) Transposable elements as genetic regulatory substrates in early development. Trends Cell Biol 23:218-26|