Genome organization by Klf4 regulates transcription in pluripotent stem cells.

How nuclear architecture and transcription are related has been a “black box” in cell biology. Whether spatial organization of chromatin in the nucleus directly regulates transcription, or initiation of transcription leads to recruitment of chromatin-interacting proteins to cause specific structure formation is still largely unclear, but both processes are critical to the establishment of cellular identity. Higher-order chromatin structure has been shown to be a distinct property of cells and exerts influence on gene networks through determining in cis and in trans interactions between genomic loci. How genomic organization influences cellular states, such as in somatic cell reprogramming, is of great interest and can shed light on the mechanisms involved. Furthermore, the organizers of higher-order chromatin structure also remain an interesting area to investigate, as they may hold useful insights on genome organization and potential application.

By using circular chromosome conformation capture and high-throughout sequencing (4C-seq), our study characterized interacting loci between the Oct4 distal enhancer and its target genes in mouse ESCs. The resulting Oct4 enhancer interactomes were found largely in gene-rich regions and correlated with higher than average gene expression compared with other non-interacting loci. This led to further investigation of whether these interactomes with active genomic marks may be associated with transcription. Using a RNA probe for Oct4 nascent mRNA in a triple labeling immuno-DNA fluorescent in situ hybridization (FISH) experiment, we identified Oct4 transcript present in nearly 80% of Oct4 interactomes, while only 40% of transcripts were detected in non-interacting loci. These colocalized structures also overlapped with RNAPII-S5P regions, supporting the conclusion that formation these chromatin structures were associated with local transcription.

We then focused on what factors may be involved in regulating formation of these structures. The Oct4 distal enhancer contains binding sites for Oct4, Sox2, and Klf4, the latter found to be enriched at Oct4-interacting sites by 4C-seq. Klf4 protein form foci within the nucleus, which through immuno-FISH showed Oct4 colocalized with its interacting partners within these foci while simultaneously overlapping with RNAPII-S5P staining regions, pointing to Klf4 as a major component bridging chromatin structure and transcription. Further analysis through immunoprecipitation and ChIP-qPCR of Klf4-interacting proteins identified Smc1, a component of the cohesin complex, to be recruited by Klf4 to the Oct4 enhancer. To test the importance of Klf4, we conducted knockdown experiments with shRNA to see how depletion of Klf4 may alter chromatin interactions and transcription. Surprisingly, Klf4-knockdown ES cells maintained levels of Oct4 transcript despite loss of interaction between 4C loci initially. Only continuous depletion of Klf4 led to decrease of Oct4 mRNA. On the other hand, inhibition of RNAPII transcription by α-amantin did not affect colocalization frequency, as seen in other studies; only overexpression of Klf4 increased colocalization frequency. This suggested a sequential order of events; first with Klf4 loading onto the Oct4 loci to colocalize Oct4 interacting loci and then transcription initiating, each seemingly independent events but requiring colocalization for sustained transcription. Shown previously by our lab, Klf4 binding at the Oct4 loci occurs through the zinc finger domain, suggesting zinc finger-dependent interactions between Klf4 and Oct4 enhancer target loci. Future experiments, such as performed by Deng et al., can further test the importance of these long-range contacts using artificial zinc finger constructs to induce colocalization between loci.

By characterizing Klf4 binding at the Oct4 loci and regulating long-range interactions through cohesin complex, the interesting question arises of how relevant this interaction is within the context of somatic cell reprogramming. The identified 4C-seq contacts were found to be unique to pluripotent cells as compared with other cell types, and these colocalized sites were also found at a lower frequency in pre-iPS cells, suggesting that forming these long-range contacts is critical within the reprogramming process. When partially reprogrammed cells were treated with DMNT1 inhibitor AZA, cells gained pluripotent cell contact frequencies after 1 d of treatment and fully reprogrammed after 5 d of induction, supporting the hypothesis that colocalization occurs earlier than transcription. Using a surface marker SSEA1 to track the subpopulation of cells that eventually fully reprogrammed, we found that the Klf4 binding intensity at the Oct4 locus and at other Oct4-interacting loci was elevated in SSEA1+ cells. Thus, Klf4 is an essential reprogramming factor functioning by binding at the Oct4 loci to set up the interaction network and initiating Oct4 transcription to complete the reprogramming process.

Klf4 is not the only pluripotent factor investigated recently for in trans chromatin interactions. Apostolou et al. investigated the interaction network for Nanog, an important late stage pluripotent marker in the reprogramming process, and found high correlation of Nanog-mediated chromatin structure formation with active gene marks, signifying active transcription. Further research using techniques like ChIA-PET can identify what factors are important for organizing specific gene loci contacts. As more powerful techniques like HiC begin to index chromatin interaction networks on a genome wide level, correlation with transcription data from RNA-seq experiments will help reveal how formation of chromatin structure may be one of the stochastic steps leading up to a fully reprogrammed cellular state.

Figure 1. Specific long-range interactions initiate during reprogramming by the loading of Klf4 to the Oct4 distal enhancer and it’s in trans interacting gene loci. Cohesin subunit Smc1 brings these loci together by interacting with Klf4. Oct4 and Sox2 binding along with Klf4 recruits transcription machinery components such as RNAPII to commence transcription. Depletion of Klf4 initially causes disruption of long-range interactions but does not immediately affect transcription; only after some time do transcription levels begin to decrease significantly, suggesting a tethering function of Klf4 in addition to transcription activation.